2026-04-06_brainforge_recruiting_process_review

Meeting Title: Brainforge Recruiting Process Review Date: 2026-04-06 Meeting participants: Kaela Gallagher, Uttam Kumaran

WEBVTT

1 00:07:18.420 ⇒ 00:07:19.720 Uttam Kumaran: Hello!

2 00:07:22.370 ⇒ 00:07:23.170 Uttam Kumaran: a…

3 00:07:24.900 ⇒ 00:07:25.910 Kaela Gallagher: How’s it going?

4 00:07:26.510 ⇒ 00:07:27.400 Uttam Kumaran: Good!

5 00:07:27.510 ⇒ 00:07:30.200 Uttam Kumaran: I feel like today’s great. I am, like…

6 00:07:31.480 ⇒ 00:07:39.710 Uttam Kumaran: I don’t know, I think people are good, like, I think Jasmine is doing well, I think Garrett’s, like, already well, I feel like we’re kind of seeing…

7 00:07:39.960 ⇒ 00:07:44.709 Uttam Kumaran: More senior people come in, and like, kind of, they’re able to just roll with the punches.

8 00:07:44.980 ⇒ 00:07:48.619 Uttam Kumaran: And they haven’t even, like, seen any of the AI stuff yet, so…

9 00:07:49.170 ⇒ 00:08:01.349 Uttam Kumaran: I feel good. Like, I mean, I’m sort of a… the problem with me is, like, if things are going really well, then I’m like, okay, what’s, like, not going well? But I think things are good, like…

10 00:08:02.360 ⇒ 00:08:05.010 Uttam Kumaran: Yeah, I feel like,

11 00:08:05.740 ⇒ 00:08:13.460 Uttam Kumaran: I was able, this weekend, I, like, basically… yeah, I just ran out of signal on Friday, like, right when I was gonna join the meeting, and we were…

12 00:08:13.540 ⇒ 00:08:18.510 Kaela Gallagher: Crazy. I was trying to respond to your text, and it was like, you have to send this via satellite, and I was like.

13 00:08:18.510 ⇒ 00:08:29.470 Uttam Kumaran: I was trying to get on the satellite, because it’s a feature where you could text people, and I was, like, outside the car, like, pointing my phone at the sky, like, I’m missing this meeting,

14 00:08:30.460 ⇒ 00:08:36.780 Uttam Kumaran: But, yeah, I was able to finally, like, do some… Thinking about,

15 00:08:37.559 ⇒ 00:08:40.260 Uttam Kumaran: Just, like, broader company vision, and…

16 00:08:40.590 ⇒ 00:08:45.869 Uttam Kumaran: You’ll see, like, a little bit about what our presentation is gonna be about at VixelCon.

17 00:08:46.030 ⇒ 00:08:52.099 Uttam Kumaran: And… yeah, I don’t know, I think we’re, like, kind of going through another… sort of, like.

18 00:08:52.200 ⇒ 00:08:54.920 Uttam Kumaran: rotation. I think this one…

19 00:08:55.640 ⇒ 00:09:01.869 Uttam Kumaran: significantly than the one before, like, I think kind of like what we did in December and November.

20 00:09:02.010 ⇒ 00:09:04.339 Uttam Kumaran: The business just, like, kind of, like, doubled.

21 00:09:04.540 ⇒ 00:09:05.180 Kaela Gallagher: Yeah.

22 00:09:05.180 ⇒ 00:09:17.650 Uttam Kumaran: more. And so, yeah, like, a lot of things broke, I kind of expected some things, but more importantly, I think, like, we have… gave us the opportunity with money to actually go spend on, like.

23 00:09:17.990 ⇒ 00:09:23.119 Uttam Kumaran: you know, Team in order to, like, invest back in, like, building the business.

24 00:09:23.270 ⇒ 00:09:29.499 Uttam Kumaran: And then this one, I think, for this quarter, and then additionally, what we saw last quarter was just, like.

25 00:09:29.660 ⇒ 00:09:33.649 Uttam Kumaran: I really just tried to grind out as much of this, like, platform situation as possible.

26 00:09:33.650 ⇒ 00:09:34.600 Kaela Gallagher: possible.

27 00:09:34.760 ⇒ 00:09:43.589 Uttam Kumaran: And so, now people are seeing what I saw, which is like, oh my god, I can create skills, I can do this, and they’re starting to run, which is amazing.

28 00:09:43.750 ⇒ 00:09:45.160 Uttam Kumaran: Now, I think…

29 00:09:45.750 ⇒ 00:09:50.840 Uttam Kumaran: we’re… we really have to figure out the sales problem. I think that’s the next thing to solve.

30 00:09:51.060 ⇒ 00:09:53.059 Uttam Kumaran: The delivery thing seems like…

31 00:09:53.200 ⇒ 00:10:07.639 Uttam Kumaran: it’s going to get solved, like, I… I see people’s excitement for things. Even in the last 3 weeks of me working with Pranav and Greg, like, they are completely different in the way they approach work now.

32 00:10:08.620 ⇒ 00:10:12.169 Uttam Kumaran: So I see that, like, moving on the delivery side, which is good, like…

33 00:10:12.440 ⇒ 00:10:15.930 Uttam Kumaran: In December, I didn’t… I was like, this is, like, a lot.

34 00:10:16.210 ⇒ 00:10:18.129 Uttam Kumaran: I feel like even last quarter.

35 00:10:18.230 ⇒ 00:10:21.209 Uttam Kumaran: It got better in some ways, tough in some other ways.

36 00:10:21.430 ⇒ 00:10:25.020 Uttam Kumaran: And then this quarter, I feel like that’s working, so sales is really…

37 00:10:25.380 ⇒ 00:10:28.460 Uttam Kumaran: the thing that we need to crush next. And then, like, a lot of…

38 00:10:28.870 ⇒ 00:10:32.790 Uttam Kumaran: The company needs me to start thinking about, like, okay, how do we get to, like.

39 00:10:33.110 ⇒ 00:10:39.030 Uttam Kumaran: like, how do we get into, like, the Fortune 500? How do we start to deepen a lot of our partnerships?

40 00:10:39.690 ⇒ 00:10:45.370 Uttam Kumaran: And, like, okay, can we set, like, a two-year road… like, what is, like, a two-year roadmap, right?

41 00:10:45.590 ⇒ 00:10:45.990 Kaela Gallagher: we need to raise.

42 00:10:45.990 ⇒ 00:10:51.300 Uttam Kumaran: Raise money for that, like… who do we need for that? Like, what is ultimately, like, a larger plan?

43 00:10:51.400 ⇒ 00:11:06.560 Uttam Kumaran: if you’d be surprised or not surprised, like, that’s, like, the sort of the side thing. That’s, like, 30 minutes or an hour a week. I’m like, okay, what is the bigger deal here? And so that was really helpful. So I was able to, like… I was off-grid, I was able to think a little bit about…

44 00:11:06.680 ⇒ 00:11:10.099 Uttam Kumaran: this past quarter and past two quarters, like.

45 00:11:10.100 ⇒ 00:11:10.610 Kaela Gallagher: Hmm.

46 00:11:10.610 ⇒ 00:11:13.760 Uttam Kumaran: What went well, and then where we kind of need to double down on, so…

47 00:11:13.760 ⇒ 00:11:18.449 Kaela Gallagher: Good. Seems like it was a good recharge time. And your girlfriend was able to come, too, and everything?

48 00:11:18.450 ⇒ 00:11:35.439 Uttam Kumaran: Yeah, yeah, she was there. I think she’s, like, getting used to camping. I don’t think she likes it. I mean, I also don’t think it’s, like… My dad really loves it, and so he’s so excited to, like, go to all these places and camp. And I camped my whole life, so I cook, and I feel great. It’s just, like, a little bit…

49 00:11:35.690 ⇒ 00:11:39.149 Uttam Kumaran: too uncomfortable. And I agree, I’m like, it’s kind of annoying.

50 00:11:40.360 ⇒ 00:11:41.110 Uttam Kumaran: Or, but…

51 00:11:41.110 ⇒ 00:11:42.250 Kaela Gallagher: That’s kind of the point, right?

52 00:11:42.250 ⇒ 00:12:01.209 Uttam Kumaran: Kind of the point, yeah. But then I was like, you need to kind of make camping yours. Like, if you like to read, or if you like to hike, or whatever it is, you can do that. Just figure out what that is. Because my dad will be like, I want to go for a run, or like, I want to go for, like, a 20-mile hike. I’m like, that’s, like, kind of too much. I’d like to hike to, like, the beach.

53 00:12:01.390 ⇒ 00:12:03.190 Uttam Kumaran: And then have a sandwich.

54 00:12:03.370 ⇒ 00:12:03.940 Kaela Gallagher: Yes.

55 00:12:03.940 ⇒ 00:12:06.410 Uttam Kumaran: Sack, go for a long walk, read.

56 00:12:06.880 ⇒ 00:12:07.600 Uttam Kumaran: Yeah.

57 00:12:07.670 ⇒ 00:12:08.610 Kaela Gallagher: Yeah, yeah, yeah.

58 00:12:08.610 ⇒ 00:12:15.929 Uttam Kumaran: that’s more my speed, or cook, like, I was just like, can I just cook for, like, 2 hours for everybody? That would be great. Like, I’ll be happy.

59 00:12:16.400 ⇒ 00:12:21.920 Kaela Gallagher: Yeah, yeah. My boyfriend and I have the same thing. Our most recent camping trip, we, like.

60 00:12:22.630 ⇒ 00:12:32.290 Kaela Gallagher: rented out this, like, mountain near San Diego. Like, this guy basically owns this entire, like, hill, and you need, like, a 4x4 to get up to the camp.

61 00:12:32.290 ⇒ 00:12:33.200 Uttam Kumaran: Oh, great!

62 00:12:33.200 ⇒ 00:12:42.260 Kaela Gallagher: But then up there, he’s built, like, out this really cool… like, all these really cool things. Like, you can hike through the property and, like, go to this cliffside.

63 00:12:42.260 ⇒ 00:12:43.460 Uttam Kumaran: You found that on Airbnb?

64 00:12:44.410 ⇒ 00:12:45.560 Kaela Gallagher: It’s on…

65 00:12:46.220 ⇒ 00:12:46.720 Uttam Kumaran: Or Verba.

66 00:12:46.720 ⇒ 00:12:49.760 Kaela Gallagher: What’s the camping? No, it’s a hip camp.

67 00:12:49.760 ⇒ 00:12:52.020 Uttam Kumaran: Hip cam. Hip Camp. Oh! Yes.

68 00:12:52.370 ⇒ 00:12:53.500 Kaela Gallagher: So cool!

69 00:12:53.500 ⇒ 00:12:58.370 Uttam Kumaran: Wait, you’ll have to send me it, because I want to go to San… I want to take my girlfriend to San Diego.

70 00:12:58.530 ⇒ 00:12:59.399 Kaela Gallagher: Oh my gosh!

71 00:12:59.400 ⇒ 00:13:04.360 Uttam Kumaran: He likes, like, the beach, and we haven’t got on, like, a really big beach trip, I would love to…

72 00:13:04.500 ⇒ 00:13:05.170 Uttam Kumaran: Yeah.

73 00:13:05.170 ⇒ 00:13:11.250 Kaela Gallagher: My requirement was that I, like, I just need an accessible toilet. That’s, like, my thing when we can.

74 00:13:11.250 ⇒ 00:13:11.930 Uttam Kumaran: Yeah, yeah, yeah.

75 00:13:11.930 ⇒ 00:13:15.280 Kaela Gallagher: And this guy built a toilet out on the side of this cliff.

76 00:13:15.770 ⇒ 00:13:27.900 Kaela Gallagher: We have a beautiful view, and it was so fun. But, like, my boyfriend’s the same thing, where he’s like, I want to go run, I want to go hike, and so he was just, like, running around the mountain, and I was just reading my book under the patio, like, you know?

77 00:13:27.900 ⇒ 00:13:28.770 Uttam Kumaran: Great. Yeah.

78 00:13:29.350 ⇒ 00:13:31.980 Uttam Kumaran: Yeah, you should send me that, that’d be great.

79 00:13:31.980 ⇒ 00:13:37.010 Kaela Gallagher: I will. It was the best, honestly, the best camping I’ve done. Luxury camping.

80 00:13:37.010 ⇒ 00:13:39.460 Uttam Kumaran: Yeah, no, that sounds like a nice middle ground.

81 00:13:39.460 ⇒ 00:13:43.349 Kaela Gallagher: Yeah, yeah, yeah. Yeah, definitely. Your girlfriend would appreciate.

82 00:13:43.520 ⇒ 00:13:44.300 Uttam Kumaran: Yeah.

83 00:13:47.360 ⇒ 00:13:53.329 Uttam Kumaran: Yeah, I mean, I thought, like, I saw your message, I think it’s fair, I think, like, why don’t… I wanted to today just, like.

84 00:13:53.560 ⇒ 00:14:03.730 Uttam Kumaran: I’ll just pull up the looms with you, and I’ll just screen share, and audio share, and, like, I’ll just try to vocalize what I’m seeing. I, myself, I was thinking also this weekend, I’m like.

85 00:14:03.930 ⇒ 00:14:06.370 Uttam Kumaran: what am I looking for? Like, and I’m like.

86 00:14:06.570 ⇒ 00:14:10.490 Uttam Kumaran: There’s some things that are, like, pretty… Like, check the box?

87 00:14:11.030 ⇒ 00:14:19.589 Uttam Kumaran: And then there’s some other things where I’m like, I don’t know, I sort of am, like, trying to find… trying to draw the line between things I see and them succeeding here.

88 00:14:20.530 ⇒ 00:14:28.499 Uttam Kumaran: And I don’t even know if that’s right, like, this is where I think a lot of our feedback on our process is, like, we need to DQ more. Okay, well.

89 00:14:28.500 ⇒ 00:14:28.820 Kaela Gallagher: Yeah.

90 00:14:28.820 ⇒ 00:14:40.320 Uttam Kumaran: that’s because, like, I typically… I’m, like, oftentimes being like, I see how this person could get there. And in fact, that’s, like, a lot of… one of the things that I took some notes on this weekend was, like.

91 00:14:40.540 ⇒ 00:14:50.179 Uttam Kumaran: was, like, what do we double down on versus, like, what do we stop doing? And a lot of the people issues that we had were, frankly caused by me being, like, this person

92 00:14:50.300 ⇒ 00:14:58.370 Uttam Kumaran: I think can get there one day. And I think there’s, like, sort of two things that came out of it. Like, one, you know, it’s so tough because a lot of people

93 00:14:58.510 ⇒ 00:15:00.760 Uttam Kumaran: at Brainforge today, were those people.

94 00:15:00.960 ⇒ 00:15:09.090 Uttam Kumaran: Where they didn’t have it, they didn’t have shit together when I met them, or it was like, they’re totally unqualified for the role they’re doing right now.

95 00:15:09.280 ⇒ 00:15:09.880 Uttam Kumaran: Yet!

96 00:15:09.880 ⇒ 00:15:10.400 Kaela Gallagher: What?

97 00:15:10.400 ⇒ 00:15:11.630 Uttam Kumaran: Yeah, yeah, go ahead.

98 00:15:11.630 ⇒ 00:15:18.330 Kaela Gallagher: What do you think is the quality that allowed those people to get to where they are today, and be, like, really strong team members?

99 00:15:18.330 ⇒ 00:15:26.120 Uttam Kumaran: Well, like, I think you don’t… and this is put in such a negative way, like, one is, like, there is a graveyard. I don’t know if that’s right, but there is a lot of people that didn’t.

100 00:15:26.740 ⇒ 00:15:27.180 Kaela Gallagher: Yeah.

101 00:15:27.180 ⇒ 00:15:30.489 Uttam Kumaran: And so there’s a lot of people that have come to the company and not.

102 00:15:30.830 ⇒ 00:15:41.809 Uttam Kumaran: there are some people that, yes, in the similar basket of people I found, they… they thrived, and they rolled with the punches, and they really trusted me, and I…

103 00:15:41.960 ⇒ 00:15:46.419 Uttam Kumaran: was, just like I am to everybody, super transparent. And,

104 00:15:46.850 ⇒ 00:16:03.560 Uttam Kumaran: I don’t know, it’s a good question. Like, I think Robert talks about this, too. I just feel like sometimes I hear something in the way people talk about their technical work, or… like, I… a lot of times, for example, when I interview engineers, I’m a lot less… I don’t care much about,

105 00:16:04.110 ⇒ 00:16:23.090 Uttam Kumaran: and we should actually care about that now, which is why it’s different, but, like, for example, a lot of people, I didn’t care if they even… I ever saw them, or they were… they came across professional. I was like, if… if they’re, like, working on AI in their free time, and their hobby is, like… or even if they have a… I used to ask people, like, what is your hobby? Or what is your passion outside of work? And…

106 00:16:23.130 ⇒ 00:16:31.560 Uttam Kumaran: when people are able to go deep on something and sort of, like, start to ramble, I’m like, okay, they’re… they are a passionate person, and I can… I can show them

107 00:16:31.770 ⇒ 00:16:41.939 Uttam Kumaran: how to become passionate about this. Like, I can give clear, and I can show… I can… I can get them wins and start to get them down. So that’s a lot of what I look for.

108 00:16:42.710 ⇒ 00:16:55.329 Uttam Kumaran: But then again, you make the mistake sometimes of some engineers, they’re so focused on just engineering that they, like, don’t see that we’re delivering for a client, and it doesn’t matter whether it’s perfect, it’s that the outcome is hit.

109 00:16:55.930 ⇒ 00:17:00.820 Uttam Kumaran: You know, so I don’t know, like, maybe you’ll notice as I watch some of these, like, some of the…

110 00:17:01.420 ⇒ 00:17:09.040 Uttam Kumaran: Sort of, you know, intangibles, but, like, yeah, I also am not, I’m not, like, 100% sure.

111 00:17:09.480 ⇒ 00:17:13.469 Uttam Kumaran: Which is hard, which is, like, either… either we should…

112 00:17:14.030 ⇒ 00:17:18.170 Uttam Kumaran: be like, okay, that’s, like, my intuition’s totally out the door, and, like, I just come in at the end.

113 00:17:18.760 ⇒ 00:17:19.250 Kaela Gallagher: It’s like…

114 00:17:19.250 ⇒ 00:17:22.699 Uttam Kumaran: somewhat… and sort of, like, one thing that I, like,

115 00:17:23.530 ⇒ 00:17:30.230 Uttam Kumaran: I wrote down… let me just pull this up… I wrote down, like, Where is this?

116 00:17:54.860 ⇒ 00:17:59.320 Uttam Kumaran: Yeah, okay, so… Ho.

117 00:18:00.380 ⇒ 00:18:02.709 Uttam Kumaran: It’s not even here.

118 00:18:06.630 ⇒ 00:18:09.370 Uttam Kumaran: Oh, and I sent this to Robert yesterday.

119 00:18:36.720 ⇒ 00:18:37.860 Uttam Kumaran: Okay.

120 00:18:43.790 ⇒ 00:18:48.579 Uttam Kumaran: Yeah, so, I mean, I think one of the pieces that I found was I was like.

121 00:18:48.860 ⇒ 00:18:51.269 Uttam Kumaran: We hired several people that didn’t work out.

122 00:18:51.920 ⇒ 00:18:52.460 Kaela Gallagher: No.

123 00:18:52.460 ⇒ 00:19:05.749 Uttam Kumaran: I… I really dislike being, like, that person did something wrong, like, I think we set the expectations wrong, or we didn’t find that they didn’t… they weren’t tuned to the job fast enough. And turning that aside, like, turning that one other way.

124 00:19:06.310 ⇒ 00:19:15.079 Uttam Kumaran: one, I think, like, I’m sort of thinking about, like, there has to be two rules. One is the person that we’re selecting for the job, or to interview, like.

125 00:19:15.310 ⇒ 00:19:19.909 Uttam Kumaran: they need to be quali- like, if they make it through, they need to be qualified to hit the ground running. I think Jasmine.

126 00:19:19.910 ⇒ 00:19:20.320 Kaela Gallagher: Yeah.

127 00:19:20.540 ⇒ 00:19:22.810 Uttam Kumaran: See that? Or, if they’re not.

128 00:19:23.040 ⇒ 00:19:28.989 Uttam Kumaran: Then we… basically, within 30 days, ideally 14 days, need to find… find out.

129 00:19:29.680 ⇒ 00:19:30.640 Kaela Gallagher: Yes.

130 00:19:30.640 ⇒ 00:19:39.299 Uttam Kumaran: So, like, for a lot of folks, like, you… I don’t think you met Lauren, who was trying to be in operations, but Shaishu, Luke, like, I think we just…

131 00:19:39.300 ⇒ 00:19:40.450 Kaela Gallagher: Meet Shayshu.

132 00:19:40.450 ⇒ 00:19:42.069 Uttam Kumaran: Yeah, we just let it go too long.

133 00:19:42.180 ⇒ 00:19:50.069 Uttam Kumaran: And I think if you even trace back before Q4, most of the reasons are me. Like, I’m the one that was like.

134 00:19:51.030 ⇒ 00:19:58.879 Uttam Kumaran: I have this decisive side, and then I’m also, like, I’m like, damn, this person, like, has it in their hand, like, it’s us, it’s us, it’s us, you know? And so…

135 00:19:59.350 ⇒ 00:20:06.069 Uttam Kumaran: I think one is, like, if people aren’t able to hit the ground running, or we’re not sure, they immediately get put in a second bucket, which is, like.

136 00:20:06.370 ⇒ 00:20:09.370 Uttam Kumaran: They need to dominate in, like, the next…

137 00:20:09.520 ⇒ 00:20:15.070 Uttam Kumaran: 14 days. Like, B is a good example. He came in on the first day, I was like, okay, like, I…

138 00:20:15.680 ⇒ 00:20:16.729 Uttam Kumaran: He’s got it.

139 00:20:17.100 ⇒ 00:20:25.240 Uttam Kumaran: Right? So, like, within 14… it was super, super obvious. But before that, I was like, I don’t know, because we tried a couple people through this recruiting firm, it didn’t work out.

140 00:20:25.670 ⇒ 00:20:28.520 Uttam Kumaran: And the other… the risk, though, is that, like.

141 00:20:28.630 ⇒ 00:20:35.930 Uttam Kumaran: it was… we’ve… I would say in dealing with each of these, situations…

142 00:20:36.320 ⇒ 00:20:41.450 Uttam Kumaran: Led to, like, tons of emotional, like, Distraction.

143 00:20:41.770 ⇒ 00:20:45.960 Uttam Kumaran: both for Robert and I, and I think just for, like.

144 00:20:46.830 ⇒ 00:20:59.229 Uttam Kumaran: We’ve probably spent tens of thousands of dollars of our time just, like, trying to remedy these things, and, like, trying out another thing, or putting together a new situation, or saying, like, hey, why don’t we fit you here, like…

145 00:20:59.510 ⇒ 00:21:05.540 Uttam Kumaran: We… and it was a… it was… we shouldn’t have done that. Like, we… we… we mismanaged our time. Like, we should…

146 00:21:05.790 ⇒ 00:21:13.069 Uttam Kumaran: both of us should have looked at each other and been like, it’s not working out, but we just didn’t have a… we didn’t have, like, an agreed-upon frame, because I’ve always been running with, like.

147 00:21:13.250 ⇒ 00:21:17.749 Uttam Kumaran: Okay, these are the people we have, like, let’s just see, let’s just try, let’s just keep trying, you know?

148 00:21:17.750 ⇒ 00:21:18.310 Kaela Gallagher: Yeah, I am.

149 00:21:18.310 ⇒ 00:21:22.159 Uttam Kumaran: It’s like, now that you’re here, and we have a process.

150 00:21:22.570 ⇒ 00:21:25.320 Uttam Kumaran: like, I’m much more open to being like.

151 00:21:25.750 ⇒ 00:21:28.449 Uttam Kumaran: Let’s have a clear framework for how these things are gonna go.

152 00:21:28.850 ⇒ 00:21:30.139 Kaela Gallagher: Yeah, okay.

153 00:21:31.140 ⇒ 00:21:36.819 Kaela Gallagher: Okay, yeah, good to keep all that in mind, even starting at these… these screening calls, like.

154 00:21:37.200 ⇒ 00:21:42.170 Kaela Gallagher: If these candidates are showing similar traits to what maybe didn’t work out in the past.

155 00:21:42.460 ⇒ 00:21:45.340 Uttam Kumaran: Yeah, and then the other piece there is, I think, like.

156 00:21:45.500 ⇒ 00:21:52.399 Uttam Kumaran: and this is where maybe I’m interested in your take. I feel like another way we could do this is, like, at the end of every interview.

157 00:21:52.510 ⇒ 00:21:53.570 Uttam Kumaran: at the end of, like.

158 00:21:53.820 ⇒ 00:21:59.309 Uttam Kumaran: A candidate gets the final round. One person has to be… has to, like, vouch for them.

159 00:22:00.270 ⇒ 00:22:06.809 Uttam Kumaran: And be like, like, what do you think about that? I sort of am on… I said that, and I’m like, okay.

160 00:22:06.940 ⇒ 00:22:09.429 Uttam Kumaran: On one hand, I’m like, that would be amazing.

161 00:22:09.750 ⇒ 00:22:12.100 Uttam Kumaran: But then, on another hand, like.

162 00:22:13.410 ⇒ 00:22:21.699 Uttam Kumaran: when in the past has this… has anyone ever vouched for anyone that strongly at Brain4? Yeah. I’m curious if I… if… what do you think about that?

163 00:22:22.060 ⇒ 00:22:32.600 Kaela Gallagher: I think that we have some very, very tough critics in our interview processes, specifically the data one, Pranav and the AI one.

164 00:22:33.110 ⇒ 00:22:33.660 Kaela Gallagher: And…

165 00:22:33.660 ⇒ 00:22:39.189 Uttam Kumaran: So what do you, what do you think they’re critiquing tough? Like, what, what, what, like, factor is it?

166 00:22:39.750 ⇒ 00:22:44.840 Kaela Gallagher: I mean, if you look at, like, Oasis scores of an interview.

167 00:22:46.130 ⇒ 00:22:52.120 Kaela Gallagher: it is very hard to come by a 5 in any category. Sure. You know? Like, he’s just, like, he just…

168 00:22:52.120 ⇒ 00:22:54.420 Uttam Kumaran: But do you… is that a feature or a bug?

169 00:22:56.340 ⇒ 00:23:02.220 Kaela Gallagher: I think it’s… I think it’s okay, as long as he’s doing it consistently, like, each interview.

170 00:23:02.220 ⇒ 00:23:02.930 Uttam Kumaran: Yes.

171 00:23:03.450 ⇒ 00:23:11.779 Kaela Gallagher: Right? And then, if we get to a point where there’s, like, more than one person doing first rounds, and then maybe the other person that’s doing it is, like.

172 00:23:12.480 ⇒ 00:23:17.810 Kaela Gallagher: more of a Sam-style interview, then we could run into issues, because we’re grading two different.

173 00:23:17.810 ⇒ 00:23:18.370 Uttam Kumaran: Yes.

174 00:23:18.370 ⇒ 00:23:33.960 Kaela Gallagher: in the first round. But for now, I don’t think it’s that big of an issue. However, I do think that, like, first and second round interviewers, if they are passing somebody to the next round, they are vouching for them in all five categories that they were.

175 00:23:34.670 ⇒ 00:23:44.429 Uttam Kumaran: So I don’t… see, I don’t know if… I don’t think… then… I don’t think we’ve made that clear, or… and it’s not… I don’t think it’s your fault, I just think maybe we need to double down on that.

176 00:23:44.690 ⇒ 00:23:45.589 Uttam Kumaran: Which is.

177 00:23:46.460 ⇒ 00:23:50.700 Kaela Gallagher: Like, if you… if you are a first or second round interviewer, and you move somebody forward…

178 00:23:50.700 ⇒ 00:23:52.890 Uttam Kumaran: I think this person can make it to our team, yeah.

179 00:23:52.890 ⇒ 00:23:59.890 Kaela Gallagher: Yes! Like, you are saying, yes, I want this person on the team, and yes, I want this person to take 45 minutes of Utama Roberts’ time.

180 00:23:59.890 ⇒ 00:24:00.330 Uttam Kumaran: Yeah.

181 00:24:00.330 ⇒ 00:24:04.419 Kaela Gallagher: That’s what you’re saying, like… Yeah, they need to be excited.

182 00:24:04.420 ⇒ 00:24:12.740 Uttam Kumaran: Then… so then what I… then I think you should look at the first-round interviewers, and be clear with them, and be like, do you want that responsibility or not?

183 00:24:13.520 ⇒ 00:24:14.290 Kaela Gallagher: Yeah.

184 00:24:15.370 ⇒ 00:24:22.530 Kaela Gallagher: I mean, I think Awish… grades really tough. I would say Amber and Sam grade less tough.

185 00:24:22.850 ⇒ 00:24:23.890 Uttam Kumaran: I, I agree.

186 00:24:24.490 ⇒ 00:24:28.179 Uttam Kumaran: I don’t think Amber and Sam are good at interviewing, like, that well.

187 00:24:28.750 ⇒ 00:24:31.339 Uttam Kumaran: Like, I think… I think they…

188 00:24:31.550 ⇒ 00:24:37.480 Uttam Kumaran: Meaning, and not good at interviewing, I think Amber, I just think, doesn’t have experience, broadly.

189 00:24:37.940 ⇒ 00:24:39.690 Kaela Gallagher: I’m gonna replace her with Jasmine.

190 00:24:39.690 ⇒ 00:24:40.010 Uttam Kumaran: It’s.

191 00:24:40.010 ⇒ 00:24:41.390 Kaela Gallagher: Like, this week, yeah.

192 00:24:41.390 ⇒ 00:24:44.240 Uttam Kumaran: I think that’s one thing. I think on Sam, he’s just too kind.

193 00:24:44.440 ⇒ 00:24:49.260 Uttam Kumaran: And… he’s not, like…

194 00:24:49.380 ⇒ 00:24:57.290 Uttam Kumaran: You can easily get run over as an interviewer, and you have to try to build… you have to try to build a case through questioning, right?

195 00:24:57.790 ⇒ 00:24:58.500 Kaela Gallagher: Yeah.

196 00:24:58.500 ⇒ 00:25:09.649 Uttam Kumaran: It’s kind of like… I don’t know, if you watch, like, Police Body Cam, where they interview the suspect in the room at the end, they’re asking the questions to build the case.

197 00:25:09.860 ⇒ 00:25:10.410 Kaela Gallagher: Yeah.

198 00:25:10.410 ⇒ 00:25:26.749 Uttam Kumaran: They’re not asking the questions to defend you, they don’t care. They’re asking just the next question that cements, like, they have a through line. Okay, you actually weren’t where you said. I’m not gonna ask you where you were, I’m gonna ask you, like, who you called before, or, like, where did, like, what’d you…

199 00:25:26.750 ⇒ 00:25:27.359 Kaela Gallagher: for lunch.

200 00:25:27.760 ⇒ 00:25:33.629 Uttam Kumaran: Right? Like… to build the case, and so I… another put… another way to put it is, like.

201 00:25:33.740 ⇒ 00:25:41.169 Uttam Kumaran: Maybe the first round interview person has to build a case to you that if they submit a person, why is this person gonna make it to our company?

202 00:25:41.510 ⇒ 00:25:43.219 Uttam Kumaran: And maybe, like.

203 00:25:43.610 ⇒ 00:25:53.509 Uttam Kumaran: if whoever is in the first round, maybe that’s what they have to do. Like, I think we just have to think about something creative, because on the flip side, I’m gonna tell you that, especially in this call, I’m gonna say…

204 00:25:53.770 ⇒ 00:25:59.669 Uttam Kumaran: I think you got… I think you should just continue to increase the floodgates, but we need to have a really strict

205 00:26:00.100 ⇒ 00:26:04.799 Uttam Kumaran: top funnel, which is either gonna be the loom, your screening, or the first round.

206 00:26:05.060 ⇒ 00:26:07.310 Uttam Kumaran: Like, our edge is gonna be…

207 00:26:07.420 ⇒ 00:26:12.470 Uttam Kumaran: like, using AI and using our… the industry we’re in.

208 00:26:12.600 ⇒ 00:26:16.669 Uttam Kumaran: I actually think it’s gonna be helpful for us to, like, double or triple the amount of people we’re talking to.

209 00:26:17.460 ⇒ 00:26:20.679 Uttam Kumaran: In order to… but then we have to filter. Really hard.

210 00:26:21.120 ⇒ 00:26:22.000 Kaela Gallagher: Right?

211 00:26:22.000 ⇒ 00:26:30.049 Uttam Kumaran: Like, because I think you can… I think with the stuff we’re doing at our company, and that’s just the stuff that we’ve talked about, you’re gonna see the demand increase.

212 00:26:31.030 ⇒ 00:26:42.960 Uttam Kumaran: So, even coming to this call, I was like, I think you should continue to use the looms. In fact, I think you should use the looms more, and you should only reach out to the ones where we’re gonna… you’re like, this person is exactly it.

213 00:26:43.100 ⇒ 00:26:46.499 Uttam Kumaran: or me, you, and Robert’s time need to go purely to, like.

214 00:26:46.710 ⇒ 00:26:48.299 Uttam Kumaran: We need to sell this person.

215 00:26:48.780 ⇒ 00:26:49.440 Kaela Gallagher: Yeah.

216 00:26:49.440 ⇒ 00:26:55.119 Uttam Kumaran: Otherwise, I think the natural demand, I have a feeling, is going to be high, but then how do you use

217 00:26:55.340 ⇒ 00:26:58.989 Uttam Kumaran: The first round is a tool, the loom is a tool.

218 00:26:59.450 ⇒ 00:27:04.669 Uttam Kumaran: Something else as a tool to help you filter, so that you can basically, like, go wider.

219 00:27:05.060 ⇒ 00:27:06.750 Kaela Gallagher: Yeah. Yeah.

220 00:27:07.710 ⇒ 00:27:08.720 Kaela Gallagher: Hmm.

221 00:27:10.050 ⇒ 00:27:16.840 Kaela Gallagher: Yeah, I think these are really good topics for us to talk about on our… Retro next week.

222 00:27:17.220 ⇒ 00:27:23.129 Kaela Gallagher: I’m hoping by then I can get Jasmine, like, at least one interview, too, so she…

223 00:27:23.660 ⇒ 00:27:26.690 Kaela Gallagher: Like, is kind of familiar with the process.

224 00:27:27.020 ⇒ 00:27:30.290 Kaela Gallagher: Yeah.

225 00:27:30.670 ⇒ 00:27:33.630 Kaela Gallagher: I think… I think this is interesting.

226 00:27:35.640 ⇒ 00:27:42.490 Uttam Kumaran: Because what I… I’m gonna… I… I… my… my feedback to you is, like, the more you can codify what makes a great

227 00:27:42.930 ⇒ 00:27:49.930 Uttam Kumaran: Brainforge team member, the faster you can scale that… that… that, like, scorecard out.

228 00:27:49.930 ⇒ 00:27:50.830 Kaela Gallagher: Yeah.

229 00:27:50.870 ⇒ 00:27:55.200 Uttam Kumaran: like, that could… and then you can edit the Loom question, so it just answers stuff.

230 00:27:55.800 ⇒ 00:27:58.409 Uttam Kumaran: That you can literally, like, be like, yeah, and so…

231 00:27:58.630 ⇒ 00:28:04.400 Uttam Kumaran: But what I’m gonna tell you is the stuff we’re doing here, there’s a lot of demand for people to want to work at these types of companies.

232 00:28:04.400 ⇒ 00:28:04.990 Kaela Gallagher: Yeah.

233 00:28:04.990 ⇒ 00:28:06.100 Uttam Kumaran: And… and…

234 00:28:06.250 ⇒ 00:28:13.600 Uttam Kumaran: And I think our advantage in recruiting is the fact that we can think about a system where you can…

235 00:28:13.840 ⇒ 00:28:19.009 Uttam Kumaran: You can… you can have 500 active applicants filter to the two that, like, matter.

236 00:28:20.380 ⇒ 00:28:23.499 Uttam Kumaran: Versus, you have 100, and then we let in 20.

237 00:28:23.670 ⇒ 00:28:34.610 Uttam Kumaran: like, I would rather you go for 500, and we figure out, okay, what would it look like? Like, how many people do we have in process right now? Like, 30 or 50? From, let’s say, Loom.

238 00:28:34.840 ⇒ 00:28:38.150 Uttam Kumaran: you’re talking to them, you’ve reached out to them, Loom or beyond.

239 00:28:38.150 ⇒ 00:28:42.240 Kaela Gallagher: Loomer Beyond, we’re probably at… 20.

240 00:28:43.090 ⇒ 00:28:46.330 Uttam Kumaran: So what if I said you have to have a, like, I would like to see 100?

241 00:28:47.750 ⇒ 00:28:50.749 Uttam Kumaran: And, like, what… I would ask you, what… what breaks?

242 00:28:50.920 ⇒ 00:28:51.680 Uttam Kumaran: Right.

243 00:28:54.410 ⇒ 00:28:57.289 Kaela Gallagher: Seeing 100, I feel like we would have…

244 00:28:57.510 ⇒ 00:29:03.350 Kaela Gallagher: too many in loom review, like, we would feel overwhelmed with the amount of looms to go through.

245 00:29:03.350 ⇒ 00:29:04.010 Uttam Kumaran: Yes.

246 00:29:04.200 ⇒ 00:29:10.820 Kaela Gallagher: And I think that… We would probably let too many into the first round.

247 00:29:11.960 ⇒ 00:29:14.310 Uttam Kumaran: So then the… then the question goes into.

248 00:29:15.740 ⇒ 00:29:22.219 Uttam Kumaran: we should never have, like, a hundred that make it past loom, or even… or you should say.

249 00:29:22.760 ⇒ 00:29:33.289 Uttam Kumaran: they shouldn’t even get a sh… how do we prevent them from getting a shot at the loom? I think the loom is good, because ultimately, you can start to use AI to help you sift through those effectively.

250 00:29:33.290 ⇒ 00:29:33.890 Kaela Gallagher: Yeah.

251 00:29:35.510 ⇒ 00:29:45.490 Uttam Kumaran: But, again, I think you’re right in that post-first round, everybody in that post-first round group, they have to all… we have to look at that and be like, I could see…

252 00:29:45.740 ⇒ 00:29:47.330 Uttam Kumaran: These people making it.

253 00:29:47.480 ⇒ 00:29:48.710 Kaela Gallagher: Yeah. Yeah.

254 00:29:48.710 ⇒ 00:29:51.159 Uttam Kumaran: And I think that’s a good… that’s one good pinch.

255 00:29:51.910 ⇒ 00:29:56.440 Uttam Kumaran: And then the second pinch is, like, can we do something pre-first round?

256 00:29:57.200 ⇒ 00:30:06.109 Uttam Kumaran: Whether it’s technical exercise, a good example is, like, there’s some companies that are like, this is the technical exercise, submit it to an AI grader.

257 00:30:06.630 ⇒ 00:30:10.800 Uttam Kumaran: And it’ll grade with a lume. So we can think about interesting ways, but, like.

258 00:30:10.970 ⇒ 00:30:15.260 Uttam Kumaran: This is where the innovation in this group is gonna be, which is, like.

259 00:30:15.750 ⇒ 00:30:22.089 Uttam Kumaran: Okay, what if… let’s say we are trying… let’s say we are scaling to, like… let’s say the same amount of people get the first round.

260 00:30:22.440 ⇒ 00:30:26.939 Uttam Kumaran: But the amount of people that are in the whatever the round before that is, like, tripled.

261 00:30:27.140 ⇒ 00:30:32.320 Kaela Gallagher: Yeah. Yeah. Then we’re just producing a higher quality going into first round. Yeah, because…

262 00:30:32.440 ⇒ 00:30:41.889 Uttam Kumaran: You may find that, like, actually, this is where, like, just the fact that they’re bad… that they’re tough reviewers doesn’t immediately tell me much, because

263 00:30:42.120 ⇒ 00:30:46.250 Uttam Kumaran: Awish may… he actually may just be holding the bar, that we said, hold the bar.

264 00:30:46.440 ⇒ 00:30:48.559 Uttam Kumaran: Right? So then I want to find out…

265 00:30:49.500 ⇒ 00:30:54.679 Uttam Kumaran: So then, if we have that, then I’m gonna ask you, do you feel like that’s the case, or do we feel like…

266 00:30:54.920 ⇒ 00:30:58.610 Uttam Kumaran: It’s actually just, we have few people, so we want to see them get in.

267 00:30:59.080 ⇒ 00:31:03.069 Uttam Kumaran: Right? And we have to sort of, like, answer both questions.

268 00:31:03.310 ⇒ 00:31:04.640 Kaela Gallagher: You’re staying?

269 00:31:04.640 ⇒ 00:31:07.220 Uttam Kumaran: Answer the question, like, is Oasis Bar the right bar?

270 00:31:07.610 ⇒ 00:31:09.650 Uttam Kumaran: And it’s, like, the bar that we should have.

271 00:31:09.780 ⇒ 00:31:13.759 Uttam Kumaran: To solve that, And then also solve the problem with, like.

272 00:31:13.980 ⇒ 00:31:20.290 Uttam Kumaran: That person who lets you in, you’re gonna be directly responsible if that person doesn’t make it the end.

273 00:31:21.370 ⇒ 00:31:25.800 Uttam Kumaran: And, like, that’s… like, we should go back and look at all the people that didn’t make it. Why did we…

274 00:31:26.510 ⇒ 00:31:29.660 Uttam Kumaran: Why did the first person not catch that, you know?

275 00:31:29.960 ⇒ 00:31:30.610 Kaela Gallagher: Yeah.

276 00:31:31.250 ⇒ 00:31:37.849 Uttam Kumaran: So I think you should take one or two of these notch tightening things, And try it.

277 00:31:39.000 ⇒ 00:31:43.499 Uttam Kumaran: I have a couple… that’s why a couple of the things where I was like, someone has to vouch

278 00:31:44.210 ⇒ 00:31:47.310 Uttam Kumaran: It rhymes with the person in the first round.

279 00:31:47.680 ⇒ 00:31:51.629 Uttam Kumaran: basically has to be like, I can see this person getting into Brain Forge.

280 00:31:51.850 ⇒ 00:31:53.520 Uttam Kumaran: And succeeding, you know?

281 00:31:53.990 ⇒ 00:31:56.619 Uttam Kumaran: But then also, if that’s a high bar, then…

282 00:31:57.110 ⇒ 00:32:06.480 Uttam Kumaran: you’re gonna have to be able to deal with the fact that, like, not many people get it, right? That’s because that’s a ton of people. So… I don’t know. And so, like.

283 00:32:07.760 ⇒ 00:32:08.520 Uttam Kumaran: like…

284 00:32:08.660 ⇒ 00:32:18.310 Uttam Kumaran: Do you still think you can handle talking to every single one of those people, or emailing with every one of those single of those people? In an event, there’s 100, or more than 100 people in that first…

285 00:32:19.080 ⇒ 00:32:22.659 Uttam Kumaran: That’s the… Like, that’s where it’s gonna go.

286 00:32:22.950 ⇒ 00:32:23.370 Kaela Gallagher: Yeah.

287 00:32:23.370 ⇒ 00:32:24.919 Uttam Kumaran: One or two steps ahead.

288 00:32:24.920 ⇒ 00:32:29.190 Kaela Gallagher: Yeah. I mean, I’m not always going to be able to talk to…

289 00:32:29.190 ⇒ 00:32:29.780 Uttam Kumaran: Yeah.

290 00:32:29.780 ⇒ 00:32:30.770 Kaela Gallagher: Every person.

291 00:32:30.770 ⇒ 00:32:34.390 Uttam Kumaran: Oh, and I don’t think you… I don’t… I think the moment they pass the first round.

292 00:32:34.810 ⇒ 00:32:38.200 Uttam Kumaran: I think your instinct is right, we should switch to, like, basically being, like.

293 00:32:38.440 ⇒ 00:32:48.709 Uttam Kumaran: we love you, we want you in here, just check the next sets of boxes. I think that’s actually much better, because then you can throw me, and I can sell, we can throw other people, and then sell.

294 00:32:49.000 ⇒ 00:32:54.270 Uttam Kumaran: Versus at the end, It’s like… almost, like.

295 00:32:55.170 ⇒ 00:33:00.239 Uttam Kumaran: Too far, but then your first line of defense has to be so solid.

296 00:33:00.440 ⇒ 00:33:00.850 Kaela Gallagher: Yeah.

297 00:33:00.850 ⇒ 00:33:09.629 Uttam Kumaran: Like, you and your relationship with those first-round reviewers has to be really good, and you should be like, you’re not a good first-round reviewer, like, I… I don’t trust that you can…

298 00:33:09.810 ⇒ 00:33:14.020 Uttam Kumaran: I don’t trust that… there’s maybe two things. Maybe they held a high bar, but they can’t communicate it.

299 00:33:14.560 ⇒ 00:33:19.190 Uttam Kumaran: They can’t communicate why… they’re like me, where they can’t really tell you, like, what they liked or didn’t like.

300 00:33:19.190 ⇒ 00:33:19.950 Kaela Gallagher: Yeah.

301 00:33:20.240 ⇒ 00:33:24.220 Uttam Kumaran: That’s bad. Like, you don’t want that.

302 00:33:24.220 ⇒ 00:33:28.600 Kaela Gallagher: I also think, like, Better than myself.

303 00:33:28.820 ⇒ 00:33:44.289 Kaela Gallagher: like, giving feedback is going to be the next round interviewer, giving the previous round interviewer feedback. Like, if somebody makes it to round two, and they’re a fail, round two should be telling round one, hey, this is why this person failed.

304 00:33:44.470 ⇒ 00:33:52.239 Kaela Gallagher: can we… can you catch that next time? You know? And, like, same for you, like, if somebody’s coming to you in a final, and there’s a gap.

305 00:33:52.620 ⇒ 00:34:01.589 Kaela Gallagher: like, maybe a technical gap, then you’re communicating that to round two of, like, hey, there was a technical gap here, like, could you have tested that in round two?

306 00:34:02.640 ⇒ 00:34:05.539 Uttam Kumaran: So, I mean, yeah, the… I mean, I think the…

307 00:34:06.370 ⇒ 00:34:11.310 Uttam Kumaran: I think the recruiting retro that we did the other time, maybe we just do… maybe you hold that?

308 00:34:11.699 ⇒ 00:34:13.270 Uttam Kumaran: Ever so often.

309 00:34:13.870 ⇒ 00:34:15.559 Kaela Gallagher: I want to do it every month.

310 00:34:15.989 ⇒ 00:34:17.719 Uttam Kumaran: I think you should do more frequently.

311 00:34:17.719 ⇒ 00:34:18.149 Kaela Gallagher: Okay.

312 00:34:18.239 ⇒ 00:34:26.460 Uttam Kumaran: I think you should do it more frequently, because ultimately what you’re going to want to do is have those feedbacks, and then move that feedback to something that’s, like, in Slack, or…

313 00:34:27.380 ⇒ 00:34:29.679 Uttam Kumaran: Through a form. You see what I mean?

314 00:34:29.810 ⇒ 00:34:30.260 Kaela Gallagher: Like.

315 00:34:30.639 ⇒ 00:34:36.459 Uttam Kumaran: If on a call we come in and you’re like, okay, let’s talk about this person, and we walk through every single step.

316 00:34:36.580 ⇒ 00:34:38.649 Uttam Kumaran: And we do that for… we do that…

317 00:34:39.239 ⇒ 00:34:44.890 Uttam Kumaran: like, we do that this week, we do that next week, we do that week after, then you’re gonna be like, okay, after every interview.

318 00:34:45.260 ⇒ 00:34:49.770 Uttam Kumaran: Second round, if they don’t pass, you have to tag the first round and say something.

319 00:34:49.969 ⇒ 00:34:53.750 Uttam Kumaran: Or, like, you know, we can think about that, but I would say go…

320 00:34:54.120 ⇒ 00:35:00.159 Uttam Kumaran: Push… push harder on it earlier to figure it out, and then you can think about the longer-term process.

321 00:35:00.160 ⇒ 00:35:01.819 Kaela Gallagher: Versus, like…

322 00:35:01.820 ⇒ 00:35:07.020 Uttam Kumaran: let’s just do it monthly, because then… then that means we’re only gonna talk 3 times about this. I’m like…

323 00:35:07.190 ⇒ 00:35:10.569 Uttam Kumaran: I think you should solve this whole thing, like, in the next few weeks, you know?

324 00:35:10.570 ⇒ 00:35:11.870 Kaela Gallagher: Yeah, okay.

325 00:35:12.190 ⇒ 00:35:13.030 Kaela Gallagher: Okay.

326 00:35:13.210 ⇒ 00:35:14.180 Kaela Gallagher: Cool.

327 00:35:14.180 ⇒ 00:35:29.109 Uttam Kumaran: good with the… I think you’re doing great with, like, the follow-ups and Slack and everything, so people are listening to your stuff. So if you tell them, hey, you’re a second-round person, if you don’t pass someone, or if you don’t underst… if you feel like you’re not going to.

328 00:35:29.690 ⇒ 00:35:35.960 Uttam Kumaran: you can always expect the first person to be the person to answer for why this person made it. And then it takes you out of the loop.

329 00:35:36.880 ⇒ 00:35:50.870 Uttam Kumaran: Similarly, those first round people, maybe you’re really, really… you tell them, like, yo, this is, like, a serious gig, and also, we need to make it more, like, if you’re a first round, then you… here’s actually, like, we’re gonna… we’re gonna get… I’m gonna get you, like, a recruiting coach.

330 00:35:51.120 ⇒ 00:36:02.980 Uttam Kumaran: We’ll find a coach who’s just… is, like, a beast at recruiting, can give feedback to all the people about them. Like, we can do those things, and then the first round, the recruiting team basically becomes…

331 00:36:03.320 ⇒ 00:36:04.740 Uttam Kumaran: First-round people and you.

332 00:36:04.980 ⇒ 00:36:10.960 Uttam Kumaran: as, like, as, like… the… the core crew. Everyone else is sort of, like.

333 00:36:11.610 ⇒ 00:36:13.879 Uttam Kumaran: Builds on that foundation, you know?

334 00:36:13.880 ⇒ 00:36:14.650 Kaela Gallagher: Yeah.

335 00:36:14.760 ⇒ 00:36:16.510 Kaela Gallagher: Yeah, yeah, yeah. Okay.

336 00:36:17.700 ⇒ 00:36:19.010 Kaela Gallagher: Sounds good.

337 00:36:20.710 ⇒ 00:36:22.899 Uttam Kumaran: Let’s go through some of these.

338 00:36:23.100 ⇒ 00:36:23.830 Kaela Gallagher: Okay.

339 00:36:25.300 ⇒ 00:36:28.870 Kaela Gallagher: I… yeah, my next meeting’s in half an hour, so…

340 00:36:28.870 ⇒ 00:36:29.330 Uttam Kumaran: Okay.

341 00:36:29.330 ⇒ 00:36:30.090 Kaela Gallagher: Good.

342 00:36:32.760 ⇒ 00:36:37.109 Kaela Gallagher: I’m actually talking to a potential partnerships person.

343 00:36:37.650 ⇒ 00:36:38.580 Uttam Kumaran: Oh, great.

344 00:36:38.580 ⇒ 00:36:39.200 Kaela Gallagher: Yeah.

345 00:36:40.360 ⇒ 00:36:42.640 Uttam Kumaran: How are, like, the sales conversations?

346 00:36:42.700 ⇒ 00:36:44.329 Kaela Gallagher: This is gonna be my first one.

347 00:36:44.330 ⇒ 00:36:45.010 Uttam Kumaran: Oh, gray, okay.

348 00:36:45.010 ⇒ 00:36:45.630 Kaela Gallagher: Yeah.

349 00:36:45.840 ⇒ 00:36:47.279 Kaela Gallagher: So we’ll see how it goes.

350 00:36:47.280 ⇒ 00:36:49.680 Uttam Kumaran: Someone that Robert put you in touch with?

351 00:36:49.680 ⇒ 00:36:52.670 Kaela Gallagher: It was, yes, the former Amplitude.

352 00:36:52.670 ⇒ 00:36:54.459 Uttam Kumaran: Oh, Hugo. Yeah, Hugo’s really nice.

353 00:36:54.460 ⇒ 00:36:55.150 Kaela Gallagher: Yeah, yeah, yeah.

354 00:36:55.360 ⇒ 00:36:57.959 Uttam Kumaran: Yeah, he was our amplitude person, he just left.

355 00:36:58.580 ⇒ 00:36:59.740 Uttam Kumaran: I liked him.

356 00:37:00.280 ⇒ 00:37:05.059 Uttam Kumaran: Yeah, I liked him. He tried to… he was really trying to work with us.

357 00:37:05.460 ⇒ 00:37:07.730 Uttam Kumaran: within their constraints.

358 00:37:07.730 ⇒ 00:37:08.330 Kaela Gallagher: Okay.

359 00:37:09.160 ⇒ 00:37:12.329 Uttam Kumaran: I mean, that’d be all kind of awesome if he joined us, because I feel like he knows

360 00:37:12.730 ⇒ 00:37:15.520 Uttam Kumaran: It’s squarely in, like, the data world, so…

361 00:37:17.380 ⇒ 00:37:18.300 Kaela Gallagher: Awesome.

362 00:37:20.080 ⇒ 00:37:22.840 Uttam Kumaran: Okay, so let me pull up recruiting…

363 00:37:32.780 ⇒ 00:37:40.310 Kaela Gallagher: Maybe we can do… like, one of each kind of role. One data, one analytics, and one AI.

364 00:37:41.700 ⇒ 00:37:42.590 Uttam Kumaran: Yes.

365 00:37:46.500 ⇒ 00:37:47.860 Uttam Kumaran: Okay…

366 00:37:51.790 ⇒ 00:37:53.380 Uttam Kumaran: Cool.

367 00:37:56.430 ⇒ 00:38:00.110 Uttam Kumaran: So let me just, odd.

368 00:38:01.670 ⇒ 00:38:03.220 Uttam Kumaran: the…

369 00:38:06.610 ⇒ 00:38:09.870 Uttam Kumaran: What is their thing for their role? Oh, rolls here.

370 00:38:10.620 ⇒ 00:38:15.900 Kaela Gallagher: If you go to Active Interviews, that section, It’s way easier.

371 00:38:15.900 ⇒ 00:38:16.910 Uttam Kumaran: Oh, he’s one zone.

372 00:38:16.910 ⇒ 00:38:17.979 Kaela Gallagher: Just go, yeah.

373 00:38:20.220 ⇒ 00:38:27.660 Uttam Kumaran: Okay, so usually what I’ll do, let’s just start at the top, so… Interesting, it’s a YouTube link.

374 00:38:28.710 ⇒ 00:38:30.769 Uttam Kumaran: I just usually opening the side.

375 00:38:31.600 ⇒ 00:38:35.739 Uttam Kumaran: And then… Wait, can you tell me if you can hear,

376 00:38:36.970 ⇒ 00:38:38.889 Uttam Kumaran: Just make sure you can hear the audio.

377 00:38:56.950 ⇒ 00:38:59.409 Uttam Kumaran: Okay, can you hear this audio?

378 00:38:59.410 ⇒ 00:39:03.670 Audio shared by Uttam Kumaran: Hi, I’m Rhodes. Professionally, I’m a data engineer focused on building…

379 00:39:03.670 ⇒ 00:39:06.219 Uttam Kumaran: But usually I watch this, like, way faster.

380 00:39:07.500 ⇒ 00:39:11.540 Uttam Kumaran: It’s like everything in my life is, like… 3X.

381 00:39:14.920 ⇒ 00:39:19.420 Audio shared by Uttam Kumaran: Building Reliable data platforms and helping organizations turn complex data environments into systems that support real business decisions.

382 00:39:19.420 ⇒ 00:39:23.829 Uttam Kumaran: Okay, so usually I’m like, is this a real person? Is their setup, like, half-decent?

383 00:39:24.150 ⇒ 00:39:25.050 Kaela Gallagher: But… Yeah.

384 00:39:25.470 ⇒ 00:39:26.300 Uttam Kumaran: St.

385 00:39:26.850 ⇒ 00:39:29.109 Uttam Kumaran: seems prepared, then I’m like, okay.

386 00:39:29.350 ⇒ 00:39:33.040 Uttam Kumaran: Usually, I, like, will listen while I open their LinkedIn.

387 00:39:34.430 ⇒ 00:39:39.599 Audio shared by Uttam Kumaran: My experience has been in consulting environments, where I’ve helped organizations modernize legacy data systems and stabilize pipelines that teams rely on for critical reporting.

388 00:39:39.980 ⇒ 00:39:44.020 Audio shared by Uttam Kumaran: What drives me in data engineering is the challenge of turning fragmented and unreliable data into something teams can…

389 00:39:44.310 ⇒ 00:39:48.210 Uttam Kumaran: It’s funny because, the thing is, I don’t…

390 00:39:48.390 ⇒ 00:39:54.800 Uttam Kumaran: whether they use AI to prepare or not is actually no longer, like, a thing for me. Actually.

391 00:39:55.110 ⇒ 00:39:59.049 Uttam Kumaran: I feel like we should prefer that they use AI to prepare, honestly.

392 00:39:59.380 ⇒ 00:39:59.710 Kaela Gallagher: as a shot.

393 00:39:59.710 ⇒ 00:40:06.239 Uttam Kumaran: that they’re actually resourceful, but mostly what I’m… what I’m looking at here is I’m like, okay, he said

394 00:40:06.410 ⇒ 00:40:09.339 Uttam Kumaran: That he’s worked with enterprises and finance.

395 00:40:09.810 ⇒ 00:40:13.679 Uttam Kumaran: To do this sort of analytics modernization.

396 00:40:13.850 ⇒ 00:40:14.670 Uttam Kumaran: like.

397 00:40:15.480 ⇒ 00:40:24.720 Uttam Kumaran: it’s a little bit sus that he doesn’t have profile or this photo, but that’s fine. He seems like a normal dude. Then I’m looking at, like, what his experience is.

398 00:40:25.150 ⇒ 00:40:30.069 Uttam Kumaran: there’s this Kolats Capital Foundation, where he was part-time.

399 00:40:31.860 ⇒ 00:40:37.160 Uttam Kumaran: Okay, this is a little bit generic of, like, a experience. And then, okay, this is, like…

400 00:40:37.340 ⇒ 00:40:41.239 Uttam Kumaran: But it looks like both of the… okay, this is actually probably his most recent role, then.

401 00:40:42.140 ⇒ 00:40:44.389 Uttam Kumaran: That, I guess he’d probably go, yeah.

402 00:40:44.390 ⇒ 00:40:45.490 Kaela Gallagher: 3 months ago.

403 00:40:45.490 ⇒ 00:40:47.330 Uttam Kumaran: Cool, so then I’m like, okay.

404 00:40:47.620 ⇒ 00:40:53.869 Uttam Kumaran: delivered large scales, blah blah blah, migrating SAS to Spark, okay, like, this is fine.

405 00:40:54.500 ⇒ 00:40:58.990 Uttam Kumaran: we’re not really, like, at this size, and so when I look at this immediately, I’m like.

406 00:40:59.340 ⇒ 00:41:01.750 Uttam Kumaran: Okay, I wonder if he can work with…

407 00:41:02.770 ⇒ 00:41:20.399 Uttam Kumaran: it’s sort of like a, kind of like a… what with Garrett, I’m like, can these people, you know, they work on smaller businesses, or faster-moving business? Are they used to sort of, like, slow legacy, and are they interested in sort of moving faster? So, that sort of things. Other than that, like, I’m just trying to see, like, how long…

408 00:41:20.780 ⇒ 00:41:26.590 Uttam Kumaran: This person has been working in data. Seems like… Since 2018-ish.

409 00:41:27.230 ⇒ 00:41:31.240 Uttam Kumaran: Or, like, actually, I mean, actually, oh, this is probably when he was in school, so then it’s, like.

410 00:41:32.140 ⇒ 00:41:35.400 Uttam Kumaran: 22, so, okay, like, yeah, 3-4 years, okay.

411 00:41:35.610 ⇒ 00:41:43.150 Uttam Kumaran: Like… Not bad. I don’t really… I actually don’t really even look at

412 00:41:43.360 ⇒ 00:41:50.460 Uttam Kumaran: this, because if you look at… I mean, I guess I studied computer engineering, but I had nothing… I didn’t do any data work, so I really don’t care about

413 00:41:51.130 ⇒ 00:41:55.940 Uttam Kumaran: I don’t really look at GPA or any of this, frankly, unless there’s, like.

414 00:41:56.180 ⇒ 00:42:07.650 Uttam Kumaran: you’ve done a program that was somewhat related to data, or the thing that you were hiring. Like, for an AI, people are taking master’s programs and stuff. The other thing I look at, if they do have their GitHub, I am gonna go look and basically see, like.

415 00:42:08.030 ⇒ 00:42:10.049 Uttam Kumaran: Do they have, like, recent stuff?

416 00:42:10.650 ⇒ 00:42:14.020 Uttam Kumaran: Kind of, not really.

417 00:42:14.250 ⇒ 00:42:18.630 Uttam Kumaran: Which is not an immediate, like, red flag, but, like.

418 00:42:19.320 ⇒ 00:42:26.950 Uttam Kumaran: Real engineers are, like, doing hobby stuff, and, like, they’ll… they’ll really have, like, Tons of contributions.

419 00:42:27.050 ⇒ 00:42:32.240 Uttam Kumaran: So that’s another thing that I look at, like, if you go to my… profile…

420 00:42:32.560 ⇒ 00:42:35.500 Uttam Kumaran: You’ll see that, like, And this is just, like…

421 00:42:37.020 ⇒ 00:42:44.050 Uttam Kumaran: Right? I’m, like, actively running this company, still doing a bunch of stuff, and these are contributions. You can link contributions from private

422 00:42:44.610 ⇒ 00:42:45.460 Uttam Kumaran: thing, so…

423 00:42:45.990 ⇒ 00:42:50.720 Uttam Kumaran: Ultimately, if you’re… unless he’s using, like, another GitHub, I’m usually expecting

424 00:42:51.190 ⇒ 00:42:54.169 Uttam Kumaran: Great engineers to be doing stuff, especially, like.

425 00:42:54.500 ⇒ 00:42:56.940 Uttam Kumaran: with AI right now, I’m like.

426 00:42:57.350 ⇒ 00:43:00.020 Uttam Kumaran: If you’re, like, an awesome engineer, you’re probably, like, trying to

427 00:43:00.720 ⇒ 00:43:03.099 Uttam Kumaran: do some AI stuff, whether I work or personal. So…

428 00:43:03.950 ⇒ 00:43:08.500 Uttam Kumaran: It’s funny, because I’m, like, articulating the things that would take me, like, 2 seconds to, like, digest.

429 00:43:08.500 ⇒ 00:43:09.270 Kaela Gallagher: Yeah.

430 00:43:09.270 ⇒ 00:43:13.949 Uttam Kumaran: That’s sort of, like, what I’m looking at. So then, like, let’s just continue, just continue watching this thing.

431 00:43:13.950 ⇒ 00:43:18.979 Audio shared by Uttam Kumaran: actually trust. Many organizations don’t lack data, they lack reliable pipelines, consistent definitions, and systems at scale as the business grows.

432 00:43:18.980 ⇒ 00:43:26.559 Uttam Kumaran: So unless their video is, like, completely trash, or, like, it’s sort of, like, it’s such a miss, I usually will watch the whole thing.

433 00:43:26.560 ⇒ 00:43:36.560 Audio shared by Uttam Kumaran: A big part of the work I enjoy is stabilizing those environments so analysts and leadership can make decisions based on metrics they’re confident in. The impact I aim to have is helping organizations move from reactive reporting into a reliable data foundation that supports long-term decision making.

434 00:43:36.560 ⇒ 00:43:49.839 Audio shared by Uttam Kumaran: For my proudest achievement, I worked on an insurance pricing pipeline system with Sunlight Financial, where leadership was experiencing inconsistent metrics during quarterly pricing reviews. The root cause was these legacy SaaS pipelines that had different insurance product groups being transformed independently that introduced non-deterministic logic, inconsistent definitions.

435 00:43:49.840 ⇒ 00:44:03.939 Uttam Kumaran: Okay, so one, as I’m hearing, like, big words, which are great, like, when you hire a data engineer, you want them to kind of be nerdy, but you don’t want… you don’t want somebody… like, there’s some people who apply as a data engineer, and they’re, like, business analysts. They want to be data engineers.

436 00:44:04.120 ⇒ 00:44:12.069 Uttam Kumaran: Because it may pay more, or, like, it’s more technical. So, what I’m looking for is, like, non-deterministic. I’m looking for systems

437 00:44:12.230 ⇒ 00:44:14.350 Uttam Kumaran: Like, the setup of systems, like…

438 00:44:14.520 ⇒ 00:44:21.829 Uttam Kumaran: real technical jargon is usually, like, what I’m looking for. I’m actually not looking for an oversimplification.

439 00:44:21.950 ⇒ 00:44:23.689 Uttam Kumaran: That’s… that’s for, like.

440 00:44:23.900 ⇒ 00:44:29.500 Uttam Kumaran: the project manager to do, right? I’m looking for people to be like, it was this system, and this issue, and, like, things like that, you know?

441 00:44:29.500 ⇒ 00:44:35.679 Audio shared by Uttam Kumaran: and duplicated transformation steps across pipelines. My role was to rebuild the transformation logic at Spark, and introduce validation checks to identify where grain mismatched.

442 00:44:35.680 ⇒ 00:44:36.639 Uttam Kumaran: Do you have to jump, by the way?

443 00:44:36.640 ⇒ 00:44:38.600 Audio shared by Uttam Kumaran: Inconsistencies were causing metric drift.

444 00:44:38.600 ⇒ 00:44:39.270 Kaela Gallagher: I haven’t happen now.

445 00:44:39.600 ⇒ 00:44:49.399 Audio shared by Uttam Kumaran: users to standardize those definitions across product groups and made pipelines deterministic and production right. The result was a stable reporting environment where pricing metrics were consistent across teams and leadership to actually rely on those outputs during quarterly decision-making.

446 00:44:49.540 ⇒ 00:44:54.910 Audio shared by Uttam Kumaran: For problem solving, one challenging issue I faced was during a migration project with the University of Texas, where we were validating workloads only at the final output layer.

447 00:44:54.910 ⇒ 00:45:00.629 Uttam Kumaran: Okay, so it looks like this… so then I’m like, okay, I’m remember… okay, so he’s actually a consultant, so it looks like…

448 00:45:00.780 ⇒ 00:45:04.060 Uttam Kumaran: This was across probably a bunch of clients, so that’s a bonus.

449 00:45:04.060 ⇒ 00:45:05.100 Kaela Gallagher: Yeah.

450 00:45:05.100 ⇒ 00:45:07.879 Uttam Kumaran: That’s a positive. He’s worked in a consulting organization.

451 00:45:08.350 ⇒ 00:45:13.970 Uttam Kumaran: like, he’s worked on a couple different systems, so I was wrong about, like, Okay…

452 00:45:14.270 ⇒ 00:45:16.290 Uttam Kumaran: I… I was looking at this, I was like, this…

453 00:45:16.960 ⇒ 00:45:23.240 Uttam Kumaran: One thing for me is, like, I… usually I will click on it and see, like, what the company is, but… okay, that’s helpful.

454 00:45:23.240 ⇒ 00:45:31.810 Audio shared by Uttam Kumaran: to determine whether they came from logic differences in the migration or upstream data. I proposed introduce read-only access to the SAS data datasets so we could separate logic replication issues from underlying data differences.

455 00:45:31.930 ⇒ 00:45:44.859 Audio shared by Uttam Kumaran: Once we isolated those two problems, we were able to troubleshoot discrepancies much more efficiently. In that same environment, I built a dependency mapping framework that converted the workload graph into an executable order and connected migration progress to Jira. That allowed migrations to run reliably, and it made it a lot easier to diagnose issues as they appeared.

456 00:45:44.860 ⇒ 00:46:00.789 Audio shared by Uttam Kumaran: One of the lessons from this experience was the importance of designing systems so problems are diagnosable, not just fixed one at a time. Looking ahead to the next year, I’m focused on strengthening my experience with modern data modeling and orchestration workloads, particularly dbt-style workloads and workflows, transformation layers, and governance practices. I spend a lot of time building large-scale spark transformation, and I want to continue improving how testing frameworks, documentation.

457 00:46:00.790 ⇒ 00:46:03.129 Uttam Kumaran: Yeah, I love it, this guy’s good, he should go, like…

458 00:46:03.430 ⇒ 00:46:07.690 Uttam Kumaran: So, what changed in the beginning is I was like, okay, resume-wise.

459 00:46:08.020 ⇒ 00:46:14.340 Uttam Kumaran: Like, he doesn’t write any of this stuff on his resume, so… Yeah. Good. Which…

460 00:46:14.840 ⇒ 00:46:17.070 Uttam Kumaran: Maybe a good thing for us.

461 00:46:17.220 ⇒ 00:46:20.530 Uttam Kumaran: But in the last, like, minute and a half.

462 00:46:20.860 ⇒ 00:46:35.819 Uttam Kumaran: he talked to me about the types of workloads he’s running, he said Spark, I heard the technical things that I kind of needed to hear. I hear that he has interest in going deeper in both learning a little bit about dbt, but continuing as a data engineer, like.

463 00:46:36.750 ⇒ 00:46:39.539 Uttam Kumaran: Yeah, I feel like he… this guy seems great.

464 00:46:40.080 ⇒ 00:46:51.039 Uttam Kumaran: But that’s just… this is, again, like, I think… I’m just showing you, like, this is… this would be a pass. I don’t know if I have anything else, though, to go off of at this point, because ultimately, if he wrote the whole thing with AI,

465 00:46:51.220 ⇒ 00:46:53.679 Uttam Kumaran: And he really doesn’t know Spark.

466 00:46:54.860 ⇒ 00:46:59.760 Uttam Kumaran: they’ll have to figure this out in the first round, like, you know, because I can only judge based on this, but…

467 00:46:59.870 ⇒ 00:47:08.259 Uttam Kumaran: He’s coming on camera, like, he’s using Zoom, Like… I feel like he…

468 00:47:08.410 ⇒ 00:47:21.009 Uttam Kumaran: he… like, he was reading stuff. I mean, I’m playing on fast mode, but, like, I can see… I think he actually knows his stuff, and looks like she just probably used AI to, like, write it out. Yeah. So I feel good about this candidate.

469 00:47:21.010 ⇒ 00:47:29.000 Kaela Gallagher: Do you feel like he would be stronger fit for a data role or an analytics role? Or maybe that’s something we have a way to determine…

470 00:47:29.000 ⇒ 00:47:34.249 Uttam Kumaran: data role, like a data engineering role, for sure. Reason being is, like, all of this stuff.

471 00:47:34.360 ⇒ 00:47:37.130 Uttam Kumaran: It is all, like, data engineering focused.

472 00:47:38.780 ⇒ 00:47:42.289 Kaela Gallagher: But then his previous roles are all analytics-focused.

473 00:47:44.250 ⇒ 00:47:49.830 Uttam Kumaran: true, but if you look at my background as well, it’s like that. So, as a data…

474 00:47:50.100 ⇒ 00:47:53.820 Uttam Kumaran: You always get into data life as this.

475 00:47:54.460 ⇒ 00:48:03.439 Uttam Kumaran: like, basically, like, a data intern or BI, and then you get a choice. Either get… go more technical, or you go towards, like, a Jasmine or Robert.

476 00:48:03.860 ⇒ 00:48:04.420 Kaela Gallagher: Yeah, yeah.

477 00:48:04.420 ⇒ 00:48:07.370 Uttam Kumaran: So that’s… so people usually just pick one or the other.

478 00:48:09.640 ⇒ 00:48:15.429 Uttam Kumaran: So, ideally, like, if he stayed here, you should see him becoming, like, senior data analyst, like.

479 00:48:15.680 ⇒ 00:48:18.120 Uttam Kumaran: Things like that, versus it looks like he just went

480 00:48:19.020 ⇒ 00:48:23.909 Uttam Kumaran: He went down the rabbit hole of, like, even some of the stuff as a data analyst here.

481 00:48:24.030 ⇒ 00:48:35.860 Uttam Kumaran: he’s implementing data warehouses, automating pipelines, like, this is beyond. So if you look at my background, too, I was like that. I started as, like, just a BI person, and then I went really deep on data engineering as well, so…

482 00:48:35.860 ⇒ 00:48:39.229 Kaela Gallagher: Okay. Yeah. Okay, cool. We’ll move them forward.

483 00:48:41.150 ⇒ 00:48:42.919 Uttam Kumaran: Cool. Let’s do another one.

484 00:48:43.160 ⇒ 00:48:45.350 Kaela Gallagher: Maybe… Martin Keller?

485 00:48:59.360 ⇒ 00:49:04.099 Uttam Kumaran: Okay, so then the… if the Loom doesn’t… if the LinkedIn doesn’t exist, then usually…

486 00:49:05.690 ⇒ 00:49:07.620 Uttam Kumaran: I’m, like, searching for him.

487 00:49:09.090 ⇒ 00:49:16.000 Uttam Kumaran: I don’t know if, like… Unlimited.

488 00:49:16.000 ⇒ 00:49:19.700 Kaela Gallagher: So what I’ve had happen is…

489 00:49:20.480 ⇒ 00:49:27.990 Kaela Gallagher: There are candidates that are paying agencies to submit applications for them.

490 00:49:28.370 ⇒ 00:49:33.460 Kaela Gallagher: And they all have a resume that looks almost exactly the same.

491 00:49:33.460 ⇒ 00:49:34.510 Uttam Kumaran: Whoa.

492 00:49:35.090 ⇒ 00:49:35.500 Kaela Gallagher: And…

493 00:49:35.500 ⇒ 00:49:36.020 Uttam Kumaran: Really?

494 00:49:36.020 ⇒ 00:49:38.450 Kaela Gallagher: Their LinkedIns never work.

495 00:49:39.250 ⇒ 00:49:40.940 Uttam Kumaran: Oh, then this is sus.

496 00:49:41.350 ⇒ 00:49:43.800 Uttam Kumaran: Oh, I didn’t know that. I thought, like, maybe…

497 00:49:44.330 ⇒ 00:49:51.020 Uttam Kumaran: you know, you just copy the wrong one or something, because I usually look for them, and I’m like, whatever. Yeah. So, but do you know… do you know what agency it is?

498 00:49:51.420 ⇒ 00:50:01.080 Kaela Gallagher: I do not. I was getting spammed with emails from them, because they had gotten ahold of my email, and

499 00:50:01.810 ⇒ 00:50:06.549 Kaela Gallagher: I had talked to a couple of them, actually, like, over a live video.

500 00:50:06.820 ⇒ 00:50:09.459 Kaela Gallagher: And they seemed fairly normal.

501 00:50:10.210 ⇒ 00:50:11.660 Uttam Kumaran: Oh, the candidates.

502 00:50:11.660 ⇒ 00:50:21.100 Kaela Gallagher: And I was like, why does your… why does your LinkedIn not work? And they’re like, oh, it’s inactive. Like, the agency I’m with, like, must have put it in there.

503 00:50:21.850 ⇒ 00:50:23.589 Kaela Gallagher: So it’s, like, really weird.

504 00:50:23.590 ⇒ 00:50:25.460 Uttam Kumaran: But what is the agency they’re with?

505 00:50:25.460 ⇒ 00:50:28.239 Kaela Gallagher: I have no idea. They’re, like, paying… they’re paying.

506 00:50:28.240 ⇒ 00:50:28.740 Uttam Kumaran: Oh.

507 00:50:28.740 ⇒ 00:50:29.310 Kaela Gallagher: me.

508 00:50:30.890 ⇒ 00:50:31.710 Uttam Kumaran: Okay.

509 00:50:31.710 ⇒ 00:50:37.819 Kaela Gallagher: Here’s Martin’s resume that he submitted. It’s hard to view from that first screen, but I tracked it down.

510 00:50:38.270 ⇒ 00:50:39.270 Uttam Kumaran: Oh, it’s in Slack.

511 00:50:40.050 ⇒ 00:50:41.729 Kaela Gallagher: I just dropped it in the chat.

512 00:50:48.200 ⇒ 00:50:54.580 Kaela Gallagher: So then I’ll look up, like, Martin Keller, and then Infuse, which is, like, his most recent role.

513 00:50:54.580 ⇒ 00:50:55.859 Uttam Kumaran: Yeah, that’s what I would do.

514 00:51:00.860 ⇒ 00:51:06.899 Kaela Gallagher: I’m seeing… him on the Himalayas app for remote workers.

515 00:51:20.120 ⇒ 00:51:22.159 Kaela Gallagher: But I’m not seeing LinkedIn.

516 00:51:24.760 ⇒ 00:51:27.389 Uttam Kumaran: I… Infuse, yeah, so… oh.

517 00:51:28.040 ⇒ 00:51:30.629 Uttam Kumaran: Yeah, so I see this is another job site.

518 00:51:31.230 ⇒ 00:51:33.529 Uttam Kumaran: Martin Keller, software engineer.

519 00:51:33.770 ⇒ 00:51:35.509 Uttam Kumaran: This is not that guy.

520 00:51:40.470 ⇒ 00:51:42.540 Kaela Gallagher: Have you heard of the Himalayas app?

521 00:51:44.260 ⇒ 00:51:44.940 Uttam Kumaran: No.

522 00:51:45.300 ⇒ 00:51:50.569 Uttam Kumaran: There’s hundreds of these, like, remote. I know about a bunch of them, but…

523 00:51:50.570 ⇒ 00:51:51.490 Kaela Gallagher: Yeah.

524 00:51:52.720 ⇒ 00:51:57.750 Uttam Kumaran: Okay, then, like, then I would look at his GitHub, I mean…

525 00:51:59.430 ⇒ 00:52:04.160 Uttam Kumaran: Still not a great story, but, like… He has stuff on there.

526 00:52:04.760 ⇒ 00:52:08.020 Uttam Kumaran: That, like, looks… Kind of legit, like…

527 00:52:08.800 ⇒ 00:52:12.120 Uttam Kumaran: For example, one, trying to suss out, like, is this guy an engineer?

528 00:52:12.380 ⇒ 00:52:13.880 Uttam Kumaran: And then two, I’m like…

529 00:52:17.150 ⇒ 00:52:24.370 Uttam Kumaran: Does he know, like, what he’s talking about? So, there’s some stuff here that’s, like, some backends…

530 00:52:24.820 ⇒ 00:52:27.790 Uttam Kumaran: These were 2 years ago, but, you know, you never know.

531 00:52:28.310 ⇒ 00:52:29.970 Uttam Kumaran: So, I don’t know.

532 00:52:30.890 ⇒ 00:52:36.180 Uttam Kumaran: give them the benefit of the doubt. I think it’s worth asking, like, yo, what is this agency? Or, like, especially if you’re getting a bunch of them.

533 00:52:36.460 ⇒ 00:52:37.390 Kaela Gallagher: But…

534 00:52:37.390 ⇒ 00:52:38.900 Uttam Kumaran: Just see what he says.

535 00:52:38.900 ⇒ 00:52:56.929 Audio shared by Uttam Kumaran: Okay, I’m Martin, I’m making this video for OpenForge, and there were some questions for this, and first of all, the motivation. Okay, so what really excites me most about working in AI is the ability to turn unsearched data into real business value and real-world applications. So, over the past few years, especially with Royal LMs and Gen AI, I’ve moved from… the world has moved from pure prediction systems to interactive, user-facing intelligence, and that shift is incredibly exciting to me.

536 00:52:56.930 ⇒ 00:53:21.240 Audio shared by Uttam Kumaran: So I’ve seen our role evolving from just building models to, getting owned end-to-end AI products. So things like REAC systems, evaluation pipelines, and user-centric AI features. So I’m particularly interested in making LLMs more reliable, explainable, and production-ready. That’s why this role really aligns, with me. And the second question is, what is the priority achievement? And one of my proudest achievements was leading the development of an AI raven recommendation system that we built, and that really improved the user engagement

537 00:53:21.240 ⇒ 00:53:24.440 Audio shared by Uttam Kumaran: We went back around 40%, and we… that really contributed,

538 00:53:24.440 ⇒ 00:53:27.750 Uttam Kumaran: Yeah, so, so far, I haven’t heard anything about, like.

539 00:53:29.570 ⇒ 00:53:32.790 Uttam Kumaran: specifics about, like, the role. It’s almost like…

540 00:53:33.170 ⇒ 00:53:40.519 Uttam Kumaran: this is an interview where you’re pitching AI transformation to, like, a small business who’s like, oh, yeah, AI…

541 00:53:40.780 ⇒ 00:53:43.020 Uttam Kumaran: AI helps customer service.

542 00:53:43.140 ⇒ 00:53:50.679 Uttam Kumaran: with X, and we get 40% of X, like, it sounds really good, but I haven’t heard anything about, like, specific agents, specific frameworks.

543 00:53:51.200 ⇒ 00:53:51.710 Kaela Gallagher: Yeah.

544 00:53:51.710 ⇒ 00:53:56.119 Uttam Kumaran: said a couple of the right things, like evals and RAG, okay, but…

545 00:53:56.870 ⇒ 00:54:00.199 Uttam Kumaran: Like, if you… if you called anyone on our team and are like.

546 00:54:00.390 ⇒ 00:54:07.560 Uttam Kumaran: what do you like about AI? And tell me some… and this… this, again, maybe just, like, our questions need to improve, but I’m interested in, like.

547 00:54:07.780 ⇒ 00:54:11.940 Uttam Kumaran: Okay, there’s some frameworks that are really popular right now, like, are you familiar with those?

548 00:54:12.150 ⇒ 00:54:17.969 Uttam Kumaran: Like, building LLM systems is saying that, like, it’s like.

549 00:54:18.260 ⇒ 00:54:22.989 Uttam Kumaran: I work on houses, like, okay, that’s, like, so… that’s, like, too much, right?

550 00:54:23.100 ⇒ 00:54:28.110 Uttam Kumaran: So, I’m gonna keep watching, seeing is there, like, any place where he gets a little bit more technical into, like, his…

551 00:54:29.260 ⇒ 00:54:30.809 Audio shared by Uttam Kumaran: Yeah. To, try to put some revenue.

552 00:54:30.810 ⇒ 00:54:31.370 Uttam Kumaran: Yeah.

553 00:54:31.620 ⇒ 00:54:41.629 Kaela Gallagher: I was gonna say, I just dropped, a screenshot in the chat, but on his application as well, when it asks about, like, what AI tools you’re using, and how you’ve…

554 00:54:41.850 ⇒ 00:54:48.199 Kaela Gallagher: applied them and that kind of thing, like, he gives a really vague answer about, like, using chat.

555 00:54:48.200 ⇒ 00:54:51.190 Uttam Kumaran: Yeah, I use AI tools like ChatGPT and Gemini.

556 00:54:51.420 ⇒ 00:54:53.479 Uttam Kumaran: I apply AI concepts.

557 00:54:54.250 ⇒ 00:55:00.280 Uttam Kumaran: Yeah, so then at this point, I’d be like, this is done, because I… what we’re probably gonna see is that

558 00:55:00.650 ⇒ 00:55:04.070 Uttam Kumaran: and this is more intense for the AI rules.

559 00:55:04.180 ⇒ 00:55:09.940 Uttam Kumaran: is people… and this is where I was… what I’m kind of mentioning is people want to be in this field really badly.

560 00:55:12.200 ⇒ 00:55:13.700 Uttam Kumaran: Unfortunately, like.

561 00:55:14.410 ⇒ 00:55:18.799 Uttam Kumaran: at my company, you need to have come with that building system. Doesn’t mean you have to do it on the job.

562 00:55:18.970 ⇒ 00:55:22.690 Uttam Kumaran: You could be doing it on the side. Like, Mustafa was doing it on the side.

563 00:55:22.870 ⇒ 00:55:30.960 Uttam Kumaran: Dan was, like, doing on the side. But, like, you still need to be doing it legitimately, like, exploring, and so just using Gemini.

564 00:55:32.230 ⇒ 00:55:32.580 Kaela Gallagher: Yeah.

565 00:55:32.580 ⇒ 00:55:33.710 Uttam Kumaran: You know, it’s not enough.

566 00:55:34.090 ⇒ 00:55:37.299 Uttam Kumaran: So, like, if… yeah, so that would kind of be it, so…

567 00:55:37.300 ⇒ 00:55:47.380 Audio shared by Uttam Kumaran: project while scaling the model while maintaining the low latency, and ensuring the recommendations stayed relevant over time. So, I addressed this by building a robust SAML pipeline. I optimized the feature training, engineering, and introducing the continuous

568 00:55:47.720 ⇒ 00:55:48.589 Audio shared by Uttam Kumaran: This is good.

569 00:55:48.590 ⇒ 00:55:49.409 Uttam Kumaran: for machine learning.

570 00:55:50.360 ⇒ 00:55:50.860 Uttam Kumaran: This is an AI.

571 00:55:50.860 ⇒ 00:56:12.970 Audio shared by Uttam Kumaran: So the system was aligned with real user behavior. And that ultimately made a solution both technical strong and business rankful. And the third question is about problem solving. In one case, we had a model underperforming in our real-time and non-detection system for IoT devices. The issue was that the model performed well offline by struggling production. So I approached this by first analyzing the data trade and latency bottleneck, and we’re discovering consistency between training and live data. I then improved the pipeline by adding real-time data validation, better feature alignments, and monitoring tools. We also fine-tuned the

572 00:56:12.970 ⇒ 00:56:15.300 Audio shared by Uttam Kumaran: model, and introduced the fallback mechanisms, which result…

573 00:56:15.300 ⇒ 00:56:29.729 Uttam Kumaran: Yeah, see, this is also the thing, is, like, you would never say this. If you’re, like… if, like, me and Sam were talking, and you’re like, how did you improve this? I would never say, yeah, I fine-tuned the model, and then introduced a while back, and I can, like.

574 00:56:31.100 ⇒ 00:56:34.430 Uttam Kumaran: It’s like, what? Like… it’s like, that’s not how you…

575 00:56:34.600 ⇒ 00:56:38.849 Uttam Kumaran: This is not how, like, if you’re an engineer, you talk about engineer. It’s like…

576 00:56:39.130 ⇒ 00:56:49.589 Uttam Kumaran: It’s like, sort of, you’re, like, you’re reading a script, I am an engineer, or, like, I write code today, like, it’s not… that’s not… it’s just, it’s like, then these are the… these are the tells, right?

577 00:56:50.010 ⇒ 00:57:00.100 Uttam Kumaran: Like, one is, like, he’s reading… I don’t… and again, I don’t mind reading from a script, but the other guy, something was different where he actually did those Spark… I could tell he was… he did those spark things.

578 00:57:00.370 ⇒ 00:57:09.530 Uttam Kumaran: Yeah. It was… he actually, I think, again, I think he probably wrote it first, used AI to do it, but he wrote, like, these are the Spark workloads, he mentioned University of Texas, like…

579 00:57:09.680 ⇒ 00:57:18.209 Uttam Kumaran: Okay, like, I think he knows what he’s doing, you know, more than this is, like, if you said AI, write me a story about how I did this, it would write this.

580 00:57:18.500 ⇒ 00:57:20.000 Kaela Gallagher: Yeah. Okay.

581 00:57:20.000 ⇒ 00:57:23.190 Audio shared by Uttam Kumaran: Which resulted in alternative databases and data evaluation techniques.

582 00:57:23.190 ⇒ 00:57:23.590 Kaela Gallagher: probably…

583 00:57:23.590 ⇒ 00:57:29.670 Uttam Kumaran: Okay, sorry if it’s painful, but I just wanna, like… I just wanna share what I’m thinking. I’ve never, I’ve never, like…

584 00:57:30.170 ⇒ 00:57:31.910 Kaela Gallagher: I said it out loud.

585 00:57:31.910 ⇒ 00:57:34.799 Uttam Kumaran: Sorry if this is really awkward or painful.

586 00:57:34.800 ⇒ 00:57:35.960 Kaela Gallagher: It’s good.

587 00:57:35.960 ⇒ 00:57:36.280 Uttam Kumaran: Okay.

588 00:57:36.280 ⇒ 00:57:48.640 Kaela Gallagher: Okay, so we will… Skip him… okay, let’s do… Chukwudi? This is Awashi’s referral.

589 00:57:50.540 ⇒ 00:57:51.330 Uttam Kumaran: Yay.

590 00:58:07.670 ⇒ 00:58:09.900 Uttam Kumaran: Cool, I like seeing that there’s…

591 00:58:10.070 ⇒ 00:58:11.720 Uttam Kumaran: A lot of stuff up here.

592 00:58:17.220 ⇒ 00:58:18.420 Uttam Kumaran: This is great.

593 00:58:22.480 ⇒ 00:58:25.780 Uttam Kumaran: This is all great, this is exactly like what we’re doing.

594 00:58:31.720 ⇒ 00:58:32.710 Uttam Kumaran: Cool.

595 00:58:33.780 ⇒ 00:58:39.789 Uttam Kumaran: Okay, so he’s had this thing called touring, and then… oh, oh, so touring is probably some type of, like, online…

596 00:58:40.650 ⇒ 00:58:42.269 Uttam Kumaran: Training company or something?

597 00:58:44.170 ⇒ 00:58:47.129 Kaela Gallagher: Yeah, looks like he’s, like, a coach or something there.

598 00:58:51.240 ⇒ 00:58:53.939 Kaela Gallagher: Okay, so he applied for…

599 00:58:54.850 ⇒ 00:58:57.379 Uttam Kumaran: Yeah, but this doesn’t… I think this is, like…

600 00:58:57.860 ⇒ 00:58:58.850 Kaela Gallagher: Data.

601 00:58:58.850 ⇒ 00:59:02.130 Uttam Kumaran: Yeah, but I think he’s probably doing one of these. So basically.

602 00:59:02.690 ⇒ 00:59:05.369 Uttam Kumaran: There’s a lot of companies right now that are literally, like.

603 00:59:05.780 ⇒ 00:59:12.709 Uttam Kumaran: do your job on camera so we can build AI to, like, learn how to do your job. You can get paid a little bit amount of money.

604 00:59:13.040 ⇒ 00:59:16.160 Uttam Kumaran: Another version of this is this company called Mercor.

605 00:59:16.780 ⇒ 00:59:24.090 Uttam Kumaran: Like, you can basically just, like…

606 00:59:25.410 ⇒ 00:59:30.129 Uttam Kumaran: like, the dumbest version of this right now is there’s a lot of companies in the US that are building

607 00:59:31.440 ⇒ 00:59:39.160 Uttam Kumaran: robots that fold your laundry and stuff, and they’re paying people in India and the Philippines to just fold laundry on camera, and using the data to train.

608 00:59:40.020 ⇒ 00:59:42.650 Uttam Kumaran: This is, like, a little bit more of an advanced version of that.

609 00:59:42.650 ⇒ 00:59:42.990 Kaela Gallagher: Okay.

610 00:59:43.450 ⇒ 00:59:45.589 Uttam Kumaran: It’s the tangent, but…

611 00:59:45.590 ⇒ 00:59:46.819 Kaela Gallagher: No, no, okay.

612 00:59:47.210 ⇒ 00:59:47.900 Kaela Gallagher: Yeah.

613 00:59:48.150 ⇒ 00:59:54.359 Uttam Kumaran: So, like, it seems like that’s what this is, where basically you can probably go in there and, like, do some work, and they’ll probably pay you a little bit.

614 00:59:54.360 ⇒ 00:59:58.059 Kaela Gallagher: Oh, it looks like he was also a data scientist for them for 3 months.

615 00:59:58.060 ⇒ 01:00:01.649 Uttam Kumaran: Oh, okay, then maybe he was working on the platform? I don’t know.

616 01:00:03.140 ⇒ 01:00:04.240 Uttam Kumaran: Worth, I mean…

617 01:00:04.380 ⇒ 01:00:04.810 Kaela Gallagher: Okay.

618 01:00:04.810 ⇒ 01:00:10.839 Uttam Kumaran: Yeah, worth someone asking in the next round? But then I… that’s… so that’s something that I’d probably be, like, what did you do for Turing?

619 01:00:11.010 ⇒ 01:00:12.420 Uttam Kumaran: Like, why did you go back?

620 01:00:12.780 ⇒ 01:00:19.329 Uttam Kumaran: Otherwise, this stuff looks fine, and he’s, like, actively reading about

621 01:00:20.680 ⇒ 01:00:24.519 Uttam Kumaran: I mean, he has a data background, he’s actively reading about AI stuff, which is great.

622 01:00:24.730 ⇒ 01:00:25.080 Kaela Gallagher: Yeah.

623 01:00:25.080 ⇒ 01:00:30.339 Uttam Kumaran: I think the other thing I’ll probably look at is, like, I’ll just open his resume.

624 01:00:32.780 ⇒ 01:00:35.449 Uttam Kumaran: Let’s see, can I see a bigger…

625 01:00:37.040 ⇒ 01:00:40.309 Uttam Kumaran: Okay, I mean, this looks generally fine.

626 01:00:46.800 ⇒ 01:00:54.200 Uttam Kumaran: And then… I’ll look at this portfolio… Okay, I mean, it’s…

627 01:00:55.160 ⇒ 01:01:10.580 Uttam Kumaran: It’s nice to see that he’s actually, like, vibe coding stuff. I think this is where I’ll… I’m gonna look at his thing, and this is great, this is, like, what we really like to see. But the problem, what I’m seeing here is it’s exactly 3 contributions, which makes me think that,

628 01:01:11.750 ⇒ 01:01:16.019 Uttam Kumaran: There’s something on a loop that’s running, that’s just, like, contributing?

629 01:01:16.310 ⇒ 01:01:22.390 Uttam Kumaran: This is another common thing that, sometimes people do.

630 01:01:22.590 ⇒ 01:01:25.430 Uttam Kumaran: Because they know that recruiters look at this.

631 01:01:26.150 ⇒ 01:01:28.749 Uttam Kumaran: And there’s automated recruiters that look at this.

632 01:01:28.990 ⇒ 01:01:32.840 Uttam Kumaran: And so… If you just have it green, it’s, like, good.

633 01:01:33.140 ⇒ 01:01:38.230 Uttam Kumaran: Yeah. Like, cause a good reason why is you can tell, like,

634 01:01:39.280 ⇒ 01:01:41.150 Uttam Kumaran: Like, I don’t work on the weekends.

635 01:01:42.270 ⇒ 01:01:45.019 Uttam Kumaran: you know… This is a 7-day calendar.

636 01:01:46.140 ⇒ 01:01:49.909 Uttam Kumaran: So I don’t work on Sunday or Saturday that often, I mean, maybe some recently.

637 01:01:50.110 ⇒ 01:01:55.069 Uttam Kumaran: So, it’s like… You… this should be kind of empty, right? So…

638 01:01:55.690 ⇒ 01:02:05.020 Uttam Kumaran: This is just so weird, because I feel like I’ve never… these are just, like, sort of weird, small things that I’m, like, just trying to go notice. But okay, whatever. So far, whatever, let’s see.

639 01:02:05.940 ⇒ 01:02:10.380 Uttam Kumaran: Let me look at, let’s watch the video.

640 01:02:14.070 ⇒ 01:02:23.709 Audio shared by Uttam Kumaran: Okay, okay, my name is, Samuel Bikley, and I’ll say, like, my journey to data engineering has been, moving from studying the electrical engineering

641 01:02:23.710 ⇒ 01:02:31.610 Audio shared by Uttam Kumaran: and data analysis, then working as a data engineer. And I would say, like, it’s more of, like, the fact that I’ve always been hungry for more expertise than trying to, like, improve myself.

642 01:02:31.610 ⇒ 01:02:54.430 Audio shared by Uttam Kumaran: And I would say, working as a data engineer in my past show, like, ByPower has shown me more, different, foresight of life when it comes to, like, making, key insights that drive businesses. I’ve seen my insights see if companies, you know, companies that were having, like, financial deficits, they were able to recover, like, Bypower, and I’ve seen my insights be able to, like, help, like, businesses grow. So, these are the things that spoil me to keep on having that space to want to be a data engineer expert.

643 01:02:54.530 ⇒ 01:03:03.200 Audio shared by Uttam Kumaran: I’ll… about my previous movements, I was seeing my previous movements has to do with, A, my first job while I was working at Bypower, a fintech company. I helped them to cover about 65 million Naira, and .

644 01:03:03.200 ⇒ 01:03:15.319 Uttam Kumaran: Yeah, so, so far, I mean, I think it’s clear he likes what he does. I feel like… I feel like he’s gonna probably talk about some engineering. The one problem I’m having sometimes is communication.

645 01:03:15.500 ⇒ 01:03:28.449 Uttam Kumaran: Like, I think… and this happens with people that are, like, either they’re… recently came to the States, or they’re in other countries, is just, like, for us, because we’re so client-facing.

646 01:03:28.560 ⇒ 01:03:32.999 Uttam Kumaran: We really are trying to hold a higher bar for, like, can this person…

647 01:03:33.430 ⇒ 01:03:42.959 Uttam Kumaran: like, have, like, really thorough communication in English, right? And I think this is something where, prior, I cared really little, because I’m, like, I can understand what people are saying.

648 01:03:43.070 ⇒ 01:03:45.470 Uttam Kumaran: It’s more about the work anyways. Now…

649 01:03:45.770 ⇒ 01:03:48.370 Uttam Kumaran: this is where I feel like my…

650 01:03:48.500 ⇒ 01:03:52.650 Uttam Kumaran: We’ll watch a little bit more, but I feel like the biggest thing here is, like.

651 01:03:53.160 ⇒ 01:03:56.029 Uttam Kumaran: Can the next interview confirm that, like.

652 01:03:56.700 ⇒ 01:04:00.910 Uttam Kumaran: they’re very, like, almost… they’re very fluent, I don’t know what fluent, or what do they say, like.

653 01:04:02.440 ⇒ 01:04:04.849 Uttam Kumaran: fluent or natural speaking, or I don’t know, whatever.

654 01:04:04.850 ⇒ 01:04:05.239 Kaela Gallagher: Is that there is.

655 01:04:05.790 ⇒ 01:04:07.629 Uttam Kumaran: Yeah. You’re in that category.

656 01:04:07.630 ⇒ 01:04:08.390 Kaela Gallagher: Native.

657 01:04:08.700 ⇒ 01:04:13.479 Uttam Kumaran: Native, yeah. So, like, native, or, like, Fluent, or whatever the one right below native is.

658 01:04:13.710 ⇒ 01:04:14.849 Uttam Kumaran: That’s sort of it.

659 01:04:15.010 ⇒ 01:04:19.019 Uttam Kumaran: and, like, that’s probably where he’s gonna get jammed.

660 01:04:19.490 ⇒ 01:04:20.359 Uttam Kumaran: But let’s see.

661 01:04:20.360 ⇒ 01:04:44.039 Audio shared by Uttam Kumaran: It took me, a bit, to get into knowing more about the data, because I was just training there. I had to, like, collaborate with, the team, and also the developer team, finance team, just to get to understand, defense, track record, historical, and to be able to, like, uncover where the, deficits were coming from. And yeah, it was really very impactful. Another impact for, impactful impact I did there was saving the company by, helping them cut their PQA analytic cost from $6, 000 t o$ 2,500. That was really very huge for me.

662 01:04:44.100 ⇒ 01:04:45.350 Audio shared by Uttam Kumaran: I was very happy about that.

663 01:04:46.840 ⇒ 01:04:53.869 Audio shared by Uttam Kumaran: Okay, when it comes to data quality, pipeline, performance issue, I also want to attribute it to my time while I was in BiPower. Like, for example, the issues of the,

664 01:04:53.870 ⇒ 01:05:16.719 Audio shared by Uttam Kumaran: 65 million deficit app we covered. We were able to uncover, the, messy data of how the data were being, like, stored, because it was not properly stored in the right order, and, why… why also, there was some issues within the codebase that affected the, data quality. So, my uncovering with SQL query that I’ve written was able to keep a check on that and ensure those things don’t appear again. And, I also further processed with using, dbt to ensure that,

665 01:05:16.720 ⇒ 01:05:21.280 Audio shared by Uttam Kumaran: We have that ability test to, keep track of whenever this kind of, data issues occurs.

666 01:05:23.630 ⇒ 01:05:30.849 Audio shared by Uttam Kumaran: Okay, for technical and analytical skills, I’m committed to straighten. I would say, like, I want to do more on, you know,

667 01:05:30.930 ⇒ 01:05:50.229 Audio shared by Uttam Kumaran: let’s say, gigabytes of data, and trying to, like, be in that space, where I’m doing more of ML operations, I think it would be really something I want to strengthen more. I haven’t done a lot on that area, and I’m looking to, like, grow myself by working on it. Self… I’m going on self-developed projects where I’m going to make use of ML hops and build a pipeline that migrates data from a warehouse, or from a data lake into a, Postgres, so I’m looking…

668 01:05:50.230 ⇒ 01:05:58.459 Uttam Kumaran: Yeah, so, like, a couple things to contrast this from the other interviews. You can tell that he’s, like, there’s some stuttering, he’s, like, talking like we all talk.

669 01:05:58.800 ⇒ 01:05:59.190 Kaela Gallagher: Right?

670 01:05:59.190 ⇒ 01:06:01.039 Uttam Kumaran: He’s maybe reading from something he wrote.

671 01:06:01.210 ⇒ 01:06:01.900 Uttam Kumaran: But…

672 01:06:02.110 ⇒ 01:06:14.120 Uttam Kumaran: he’s kind of going, great, perfect. Like, the second piece is, like, I think he mentioned DBT, he mentioned sort of the technical things that I’m, like, basically looking for. I think it all comes down to, like, whether…

673 01:06:14.260 ⇒ 01:06:15.960 Uttam Kumaran: He can handle the communication.

674 01:06:16.890 ⇒ 01:06:20.600 Uttam Kumaran: And so this is also where, like, it’s so painful to me, because…

675 01:06:20.820 ⇒ 01:06:27.070 Uttam Kumaran: I feel like there are people that we’ve recruited in the past where I’m like.

676 01:06:27.520 ⇒ 01:06:31.639 Uttam Kumaran: Damn, if you just, like, worked on English a little bit more.

677 01:06:31.770 ⇒ 01:06:36.490 Uttam Kumaran: you could easily come and make money here. And so I’m like, what do we do? Like…

678 01:06:37.140 ⇒ 01:06:45.439 Uttam Kumaran: I’m almost like, damn, we should just have someone that’s, like, an ESL coach that’s just like, okay, you’re not that great at this, but, like.

679 01:06:45.710 ⇒ 01:06:53.819 Uttam Kumaran: 30 days, we’re gonna train you up, because he’s great, this guy seems really good at engineering, and it’s sort of… I feel so bad that,

680 01:06:54.670 ⇒ 01:06:58.229 Uttam Kumaran: I just feel bad that he’s gonna get dinged on the communication.

681 01:06:58.470 ⇒ 01:07:01.419 Uttam Kumaran: I don’t know if, like, and this is just… I don’t know if I’m just being…

682 01:07:01.730 ⇒ 01:07:06.870 Uttam Kumaran: Empathetic, but, like… It’s… it’s sort of tough, you know?

683 01:07:07.350 ⇒ 01:07:07.880 Kaela Gallagher: Yeah.

684 01:07:07.880 ⇒ 01:07:18.029 Uttam Kumaran: because we’re not a software company, meaning, like, we… all of these folks are going to interact with the customers. Maybe the engineers, like, data engineers and data modelers, less and less over time, but…

685 01:07:18.660 ⇒ 01:07:20.719 Uttam Kumaran: Like, still will happen, you know?

686 01:07:20.720 ⇒ 01:07:22.220 Kaela Gallagher: Yeah, yeah.

687 01:07:22.550 ⇒ 01:07:25.669 Kaela Gallagher: I would say, like, yeah, it’s tough.

688 01:07:25.840 ⇒ 01:07:32.570 Kaela Gallagher: We could see how he does on a live call for the first round, and call it out as something to be, like, very…

689 01:07:32.950 ⇒ 01:07:35.559 Kaela Gallagher: Critical on and focus on, but…

690 01:07:35.970 ⇒ 01:07:38.620 Kaela Gallagher: Yeah, I don’t know. I, I, I think…

691 01:07:38.890 ⇒ 01:07:42.959 Kaela Gallagher: Like, communication is probably the hardest thing to…

692 01:07:43.670 ⇒ 01:07:46.089 Kaela Gallagher: Find in an engineer, but also.

693 01:07:46.090 ⇒ 01:07:46.500 Uttam Kumaran: Yes.

694 01:07:46.500 ⇒ 01:07:48.209 Kaela Gallagher: It’s, like, really important to us.

695 01:07:48.700 ⇒ 01:07:55.689 Uttam Kumaran: Yeah, you’ll find one or… you’ll find… like, for… another thing is, like, I’m suspect of engineers that are really good at communicating.

696 01:07:56.020 ⇒ 01:08:10.770 Uttam Kumaran: Because I didn’t get good until, like, I had to do this… run a business. Or, like, I started working on managing teams. Like, I learned, I taught myself, like… I literally… I read this book called On Writing Well. It, like, literally taught me how to write, like, succinctly.

697 01:08:10.770 ⇒ 01:08:11.200 Kaela Gallagher: I’m like…

698 01:08:11.200 ⇒ 01:08:16.319 Uttam Kumaran: I’ve… But that’s not what being a great engineer optimizes for.

699 01:08:17.870 ⇒ 01:08:24.049 Uttam Kumaran: yes, I think as you’ve become a great engineer, you have to communicate, but you’re communicating with other engineers

700 01:08:24.189 ⇒ 01:08:27.379 Uttam Kumaran: So you’re sort of, like, you can speak the alien language a little bit.

701 01:08:27.899 ⇒ 01:08:28.519 Kaela Gallagher: Yeah.

702 01:08:28.979 ⇒ 01:08:39.899 Uttam Kumaran: So I’m not oftentimes, like… I’m actually looking for people that are sort of stumbling over their words, or, like, kind of, like, a little bit all over, because that’s great. I’m actually… if people are too…

703 01:08:40.119 ⇒ 01:08:46.159 Uttam Kumaran: put together. I’m like, dude, I don’t think… I think you’re a business person, which is fine, where that’s what we’re hiring.

704 01:08:46.490 ⇒ 01:08:47.779 Uttam Kumaran: But…

705 01:08:47.939 ⇒ 01:08:54.409 Uttam Kumaran: that’s sort of, like, the opposite of the business people. That’s how they can suss out if someone’s not a business person versus an engineer.

706 01:08:54.720 ⇒ 01:08:58.930 Uttam Kumaran: It’s easy for me to be like, okay, you’re not really, like, one of us, you know?

707 01:08:58.930 ⇒ 01:09:00.170 Kaela Gallagher: So…

708 01:09:00.580 ⇒ 01:09:06.250 Uttam Kumaran: I also agree, it’s definitely not hard, but I’m… for great engineers, I’m willing to look past

709 01:09:06.350 ⇒ 01:09:12.619 Uttam Kumaran: a lot of that. Like, that’s… that’s just, like, small paper cuts versus, like, Okay, if…

710 01:09:12.990 ⇒ 01:09:16.180 Uttam Kumaran: if he’s struggling on… on English.

711 01:09:16.529 ⇒ 01:09:23.199 Uttam Kumaran: And even for me, in watching and listening to this, I’m having a hard time understanding everything, like…

712 01:09:23.410 ⇒ 01:09:26.590 Uttam Kumaran: Okay, but again, if at this point.

713 01:09:27.840 ⇒ 01:09:35.219 Uttam Kumaran: and this is where I’ll push on you, if you watching this are like, he doesn’t fit the communication level, it’s not a pass.

714 01:09:35.500 ⇒ 01:09:40.710 Uttam Kumaran: then we should not pass him, and I’m fine with that. Then, ultimately, it’s maybe a question of, like.

715 01:09:41.029 ⇒ 01:09:44.700 Uttam Kumaran: For folks, they don’t hit the communication level, but they hit the engineering level.

716 01:09:45.020 ⇒ 01:09:47.350 Uttam Kumaran: Is there any way we could still…

717 01:09:48.260 ⇒ 01:09:54.629 Uttam Kumaran: leverage them, or is there any way that, like, yeah, we can… I don’t know. Like, I would prefer that, for sure.

718 01:09:55.140 ⇒ 01:09:55.660 Uttam Kumaran: But…

719 01:09:55.660 ⇒ 01:09:56.340 Kaela Gallagher: Yeah.

720 01:09:56.340 ⇒ 01:09:59.799 Uttam Kumaran: it’s tough because, like, yeah, I have a feeling that

721 01:10:00.630 ⇒ 01:10:02.649 Uttam Kumaran: I have a feeling it’s not gonna get better.

722 01:10:03.260 ⇒ 01:10:04.760 Kaela Gallagher: Yeah, yeah.

723 01:10:04.970 ⇒ 01:10:08.929 Kaela Gallagher: Okay. So for him, what is your instinct here?

724 01:10:10.990 ⇒ 01:10:15.920 Uttam Kumaran: I mean, it’s so hard. Like, I think he’s a good engineer.

725 01:10:16.710 ⇒ 01:10:21.180 Uttam Kumaran: But I think we may be in a similar situation as,

726 01:10:22.380 ⇒ 01:10:26.090 Uttam Kumaran: I forgot who else we interviewed that got to the final round.

727 01:10:26.280 ⇒ 01:10:28.309 Uttam Kumaran: But the communication just wasn’t there.

728 01:10:28.850 ⇒ 01:10:36.180 Uttam Kumaran: And I have a feeling that, like, he may pass the technical, but he may not… he may end up still failing later on communication, so…

729 01:10:36.530 ⇒ 01:10:37.340 Kaela Gallagher: Okay.

730 01:10:38.730 ⇒ 01:10:39.600 Kaela Gallagher: Hmm.

731 01:10:41.800 ⇒ 01:10:42.590 Uttam Kumaran: What do you think?

732 01:10:44.230 ⇒ 01:10:46.499 Kaela Gallagher: I think it’s fine to hold on him for now.

733 01:10:47.120 ⇒ 01:10:48.350 Kaela Gallagher: Okay.

734 01:10:48.910 ⇒ 01:10:55.900 Kaela Gallagher: I can go through and watch the rest of these today, and if there’s, like, not a single other data engineer that is good.

735 01:10:57.250 ⇒ 01:10:59.250 Kaela Gallagher: Maybe we can see how it does, but…

736 01:10:59.250 ⇒ 01:11:01.049 Uttam Kumaran: That’s… but that’s what I’m saying, is that I…

737 01:11:01.050 ⇒ 01:11:02.360 Kaela Gallagher: We don’t… yeah.

738 01:11:02.360 ⇒ 01:11:06.309 Uttam Kumaran: No, I’m just saying, like, your… the name of the game is gonna be increasing, though.

739 01:11:06.860 ⇒ 01:11:08.270 Uttam Kumaran: The top of the funnel.

740 01:11:08.480 ⇒ 01:11:09.939 Uttam Kumaran: Yeah. Because…

741 01:11:10.430 ⇒ 01:11:20.899 Uttam Kumaran: I just think it’s so hard, like, and we have a high bar. A lot of companies would just let folks in, because the bar is already high, and my… and I’m telling you, we’re gonna make it even higher.

742 01:11:22.360 ⇒ 01:11:29.690 Uttam Kumaran: the… the… Yeah, the solve is… I think the solve… one of the solves is…

743 01:11:29.990 ⇒ 01:11:34.899 Uttam Kumaran: putting more onus on that first interviewer to defend. The second solve is…

744 01:11:36.090 ⇒ 01:11:46.280 Uttam Kumaran: you get 500 candidates, you’re fine, too. Two that are great versus the 100, you maybe hit or miss. I don’t know. I just think that’s… it’s a numbers game.

745 01:11:47.280 ⇒ 01:11:49.430 Uttam Kumaran: And it’s tough, you know.

746 01:11:50.130 ⇒ 01:11:51.590 Kaela Gallagher: Yeah, agreed.

747 01:11:52.090 ⇒ 01:12:03.489 Kaela Gallagher: Okay, well, I gotta hop into this interview, but thank you for your help with that, and I’ll let you know if I run into any that I’m, like, struggling with. Maybe you can be a second set of eyes.

748 01:12:03.490 ⇒ 01:12:04.010 Uttam Kumaran: Okay, okay.

749 01:12:04.320 ⇒ 01:12:05.319 Kaela Gallagher: Alright, thank you!

750 01:12:05.320 ⇒ 01:12:05.900 Uttam Kumaran: Thank you. Bye.

751 01:12:05.900 ⇒ 01:12:06.270 Kaela Gallagher: Right.

Brainforge Knowledge

Explorer

2026-04-06_brainforge_recruiting_process_review_fe5f30b7

Graph View