Meeting Title: Friday Brainforge Demos & Retro Date: 2025-04-04 Meeting participants: Aakash Tandel, Luke Daque, Nicolas Sucari, Uttam Kumaran, Amber Lin, Demilade Agboola, Hannah Wang, Miguel De Veyra, Casie Aviles, Awaish Kumar, Caio Velasco
WEBVTT
1 00:00:42.830 ⇒ 00:00:44.000 Uttam Kumaran: Hi! Everyone.
2 00:00:46.756 ⇒ 00:00:47.273 Caio Velasco: So.
3 00:00:48.880 ⇒ 00:00:49.580 Uttam Kumaran: A
4 00:00:57.640 ⇒ 00:00:59.279 Uttam Kumaran: no arsenal gear, dude
5 00:01:01.760 ⇒ 00:01:04.750 Demilade Agboola: It says, switch it up, you know it’s not the same look every single time
6 00:01:04.750 ⇒ 00:01:09.049 Uttam Kumaran: I haven’t. I haven’t seen in a while, so I haven’t even I haven’t checked the
7 00:01:09.260 ⇒ 00:01:13.900 Uttam Kumaran: the scoreboards. What are they losing? So you’re just you’re stopped supporting
8 00:01:14.070 ⇒ 00:01:23.662 Demilade Agboola: Oh, no, no! So there was a break, an international break. So they went to pay for their countries. The players went to pay for their countries. But then we played on Tuesday, and we won 2 1. So you know, we’re still winning
9 00:01:24.792 ⇒ 00:01:26.617 Uttam Kumaran: Okay, okay.
10 00:01:28.090 ⇒ 00:01:31.660 Uttam Kumaran: The Lakers lost yesterday. So that’s tough.
11 00:01:35.110 ⇒ 00:01:41.170 Uttam Kumaran: The warriors have to fight. The warriors have to fight to make it in the playoffs this year. So let’s see.
12 00:01:42.270 ⇒ 00:01:45.960 Demilade Agboola: Well, the rockets have made the playoffs already, so it’s a good season. That’s a lot
13 00:01:45.960 ⇒ 00:01:47.380 Uttam Kumaran: I know.
14 00:01:48.220 ⇒ 00:01:53.810 Uttam Kumaran: Dude, that coach is really really good man. Really, great coach
15 00:01:54.260 ⇒ 00:01:55.000 Demilade Agboola: Yeah.
16 00:02:01.690 ⇒ 00:02:03.499 Uttam Kumaran: Okay, cool. I think we have
17 00:02:03.670 ⇒ 00:02:19.720 Uttam Kumaran: critical mass. Yeah, I think I kind of wanted to do a little bit of a a different format, and I may, you know, just pick on some people as we as we go through slides for today. But sort of the theme of
18 00:02:20.050 ⇒ 00:02:25.432 Uttam Kumaran: this week that I want to talk about is, I will go through our our typical slides where we just look at all of our clients.
19 00:02:26.081 ⇒ 00:02:53.529 Uttam Kumaran: I also want to get some input from everybody on how we felt about accomplishments, what we felt about challenges. And then I want to keep some time at the end, ideally for AI team or anyone to sort of just talk through like how people are using AI day to day. I think we’re having a lot of conversations about tools like hacks, things like that. I use AI like annoyingly.
20 00:02:53.530 ⇒ 00:03:23.519 Uttam Kumaran: you know, and it’s it’s been incredibly incredibly helpful. And you guys know that I wanna make, or you guys may not know is that I want to make sure everybody here. If you want to try some new agent you want to try chat, gpt, you have access to it. But I think you know, this week I learned that probably a lot of the stuff that I’m doing it just we just haven’t shared. And each of us, beyond using our own agents, can just try to find ways to use AI in our day to day. So I’ll kind of leave some time
21 00:03:23.670 ⇒ 00:03:33.560 Uttam Kumaran: toward the end, probably for Amber Miguel. Some of our folks that are using AI sort of share a couple of things, and then bunch anyone else can can sort of chime in
22 00:03:34.590 ⇒ 00:03:36.950 Uttam Kumaran: cool. So
23 00:03:37.854 ⇒ 00:03:48.529 Uttam Kumaran: yeah, I guess I’ll pause here again. I like looking at this every week. It feels like a nice, a nice ritual. Yeah.
24 00:03:49.260 ⇒ 00:04:06.020 Uttam Kumaran: I think we did. I think we’re we’re getting much, much better. On on all of these fronts, I think, in in particular. On the last item. You know, this week I was spending some time. I was spending some time with
25 00:04:06.540 ⇒ 00:04:34.450 Uttam Kumaran: in the slack channel, with demalade, in a way, sort of talking about like how we actually start to improve everyone’s knowledge. And one of the things that we’re sort of thinking about doing is, how do we get everyone up to date with all of our systems and offer sort of either certifications or more help, you know, in order to actually level up. And so part of this I started doing like one or 2 h of office hours this week. And I think we’re gonna try to continue that process into next week.
26 00:04:34.786 ⇒ 00:04:45.450 Uttam Kumaran: Where basically, we’re just available to ask questions. I think also, a wish is gonna start a channel where we can just ask any sort of engineering help questions. I want this to become like
27 00:04:45.730 ⇒ 00:05:07.210 Uttam Kumaran: like 50% more collaborative. We all get stuck on like one or 2 small things that I typically go and ask someone really quickly. And I want to create that not as a bug, but as a feature of our company, that if you do have a technical challenge, there’s someone in the company that can help you with so I think we’re gonna start a channel around around that this week. But
28 00:05:07.310 ⇒ 00:05:15.250 Uttam Kumaran: yeah, anyone else have any comments on how we did related to our mission or or values this week.
29 00:05:20.210 ⇒ 00:05:21.030 Uttam Kumaran: Cool?
30 00:05:21.713 ⇒ 00:05:36.930 Uttam Kumaran: Yeah, I guess! Shout outs, maybe I’ll I’ll go first.st I just want to shout out. I’ve been spending a lot of time with the marketing team this week. And last week we just did a webinar yesterday. Where I presented on
31 00:05:36.930 ⇒ 00:05:55.410 Uttam Kumaran: using data and go to market analytics. And it went really, really well, I think we got great footage from it. We had a whole series of Linkedin posts leading up to it. We’re going to have some posts coming out of it, and it’s our 1st of many that we’re doing in the next 3, 4 weeks. So
32 00:05:56.227 ⇒ 00:06:04.779 Uttam Kumaran: just want to shout out that whole team for sort of getting organized around this and sort of creating a process that we could follow. For all you know, new events.
33 00:06:05.830 ⇒ 00:06:09.230 Uttam Kumaran: Anyone else, with shout outs
34 00:06:16.187 ⇒ 00:06:21.620 Amber Lin: I’ll say that for my team. I think we had so much progress
35 00:06:21.960 ⇒ 00:06:34.930 Amber Lin: on ABC. While keeping the hours a lot shorter. I know we’re gonna talk about it later. But last week we spent 63 h on ABC.
36 00:06:35.260 ⇒ 00:06:38.409 Amber Lin: This week we spent 27.
37 00:06:39.280 ⇒ 00:06:45.029 Amber Lin: So humongous progress, and this week. The client was so happy.
38 00:06:45.300 ⇒ 00:06:54.330 Amber Lin: Scott, the client, and Scott was like this week, was a fantastic update. I’ve never heard them say that before.
39 00:06:54.330 ⇒ 00:07:01.660 Uttam Kumaran: Amber. Can you explain? Can you explain why reducing hours beyond just the fact that we spent less time like, why, that’s important.
40 00:07:02.340 ⇒ 00:07:03.850 Amber Lin: Okay? So
41 00:07:04.080 ⇒ 00:07:19.750 Amber Lin: we have 2 types of clients, one type of client that pays us hourly, one type that pays us a fixed fee. Right. So when it’s hourly the more hours you do, the more you get paid as long as it’s within the range. But when it’s a fixed rate for month.
42 00:07:20.210 ⇒ 00:07:26.250 Amber Lin: The more hours you do simple division math, the less rate you get per hour.
43 00:07:26.450 ⇒ 00:07:31.750 Amber Lin: So if we do 63 h.
44 00:07:32.110 ⇒ 00:07:37.229 Amber Lin: then on average, we’re probably getting $31 per hour.
45 00:07:37.350 ⇒ 00:07:53.279 Amber Lin: That that is, we are losing money, but if we do a lot less hours while delivering, then we get better rates, and then we’ll be able to pay everyone better because we’re actually profitable
46 00:07:54.530 ⇒ 00:08:21.899 Uttam Kumaran: Yes, could not say better. Everybody knows sort of. I’ve explained that multiple times. But it’s incredible. I think this is the last week was the 1st week we did a pure allocation meeting, where we looked at everybody’s time and again. Thanks for everybody for putting time into clock. If I. It’s the only way we can sort of measure. But this week was our 1st application of those decisions. And we’re already seeing, you know better optimizations. So that’s really really huge.
47 00:08:22.798 ⇒ 00:08:26.430 Uttam Kumaran: Anyone else with shout outs this week
48 00:08:27.380 ⇒ 00:08:31.379 Nicolas Sucari: Maybe a shout out to Casey. He was the 1st one to use linear asks
49 00:08:31.390 ⇒ 00:09:00.780 Nicolas Sucari: for a new holiday request. So thank you, Casey. We’re gonna be implementing some new, some more kind of asks that the team can can do through slack. For example, the tool request. So I’m gonna share that in the Brainforge team channel so that everyone can be aware of that and start using the linear asks to all of that kind of requests. So that we have everything organized. We are receiving a linear ticket on each of those asks, and then can like start our process on how to address all of those ones
50 00:09:03.890 ⇒ 00:09:04.600 Uttam Kumaran: Awesome.
51 00:09:05.240 ⇒ 00:09:24.170 Demilade Agboola: And then I also want to shout out, Casey. Had an issue to get some Api requests done for the Eden project and Casey came through. I just wanted him what my requirements were, and he just sent me back the response after like so like that, was very helpful.
52 00:09:26.170 ⇒ 00:09:27.250 Uttam Kumaran: Hell, yeah.
53 00:09:30.210 ⇒ 00:09:53.709 Aakash Tandel: I want to give a quick shout out to Annie. She did a great job walking. I’m on at the and Vlad. His new data analyst on the job team through their Meta base, and we were able to talk through kind of the other pieces or stack. So that was really good, I thought, and did a phenomenal job. And yeah, Robert and I, caught the stuff that was a little bit more engineering focused or stuff that was outside of any scope or
54 00:09:54.270 ⇒ 00:09:55.769 Aakash Tandel: yeah, I think it went really well.
55 00:09:58.490 ⇒ 00:09:59.390 Uttam Kumaran: Awesome.
56 00:10:02.970 ⇒ 00:10:04.190 Uttam Kumaran: Anyone else
57 00:10:06.639 ⇒ 00:10:33.659 Amber Lin: Yes, I’m back. So for pool parts we had a very, very, very urgent request. The client actually typed in the chat. I am pissed. I’m gonna I want to stop billing. And then and Luke especially went in and was like, Okay, we’re gonna do this. So some very brilliant firefighting that we had this week, and I wanted to shout out to Luke
58 00:10:36.080 ⇒ 00:11:02.989 Uttam Kumaran: Nico’s familiar with with pull parts. That’s really really great. I think you know, 1 1 of the pieces, I’ll explain, is, we have clients that somewhat have some great emotional intelligence. We have some clients that definitely struggle but our job is to solve problems even this morning. In stack Blitz. I sent a message which is like, Hey, I’m not sure this Pr is doing what you said. They said, don’t worry about that.
59 00:11:03.340 ⇒ 00:11:27.249 Uttam Kumaran: I was like, okay, great fine, you know. And so this is where again we are. We are trying to act always in the best interest of the client. We try not to get emotional. It’s hard sometimes. But we want to be looked at as true partners. Not just consultants that never say anything, not consultants that are like trying to protect our hours, or like extract more hours. It’s a delicate balance, though.
60 00:11:31.189 ⇒ 00:11:36.810 Uttam Kumaran: Great. Okay. If any other shout outs.
61 00:11:42.520 ⇒ 00:11:43.050 Uttam Kumaran: okay.
62 00:11:47.060 ⇒ 00:11:51.599 Uttam Kumaran: Oh, no way. Yeah. Wait. Go go on on mute, unmute
63 00:11:52.680 ⇒ 00:11:54.820 Hannah Wang: Hello, Hi!
64 00:11:58.160 ⇒ 00:12:02.630 Hannah Wang: Co-working in in my place. So that’s interesting.
65 00:12:03.730 ⇒ 00:12:08.499 Hannah Wang: Not wearing proper attire. He’s in his speech, so he’s gotta go away
66 00:12:09.820 ⇒ 00:12:16.377 Uttam Kumaran: Awesome. That’s so nice to see you guys. And then a couple of us will be together in la next weekend.
67 00:12:16.850 ⇒ 00:12:25.285 Uttam Kumaran: It’s we’re running the the cheapest off site of all time that we can afford right now. But I’m I’m very, very excited. Next,
68 00:12:25.780 ⇒ 00:12:42.379 Uttam Kumaran: hopefully, we’ll my goal this year is to one way or another. See everybody in person. I’ve crossed off some of the people on on this list, but I think I’m excited to hopefully go around the world and see a couple more folks.
69 00:12:43.790 ⇒ 00:12:48.720 Uttam Kumaran: great anyone else before we keep going
70 00:12:52.160 ⇒ 00:13:09.290 Uttam Kumaran: cool. Yeah, let me just give a quick update on on clients. So I think the urban stems client is actually going really well again. Big shout out to demalade. I’m sort of playing more like platform lead, and sort of picking off
71 00:13:09.340 ⇒ 00:13:26.199 Uttam Kumaran: one or 2 things where I can. But I think this client really really appreciates our knowledge, and you know how we’re building great systems. So super looking forward to this, I think we’re actually gonna be able to upsell them and and get a couple more dollars in the door. On Eden. Health.
72 00:13:26.550 ⇒ 00:13:38.610 Uttam Kumaran: I think maybe I’ll I’ll be for Eden or Javi. Maybe I’ll I’ll pick on someone on that team, maybe Aish. Do you want to give like a status on how you think Eden’s going
73 00:13:42.077 ⇒ 00:13:45.000 Awaish Kumar: I think it’s it’s going good. We
74 00:13:45.571 ⇒ 00:13:50.209 Awaish Kumar: like the in the today’s call. We got a good feedback on like the
75 00:13:50.650 ⇒ 00:13:53.850 Awaish Kumar: the Josh was happy with the data quality and
76 00:13:54.310 ⇒ 00:13:56.420 Awaish Kumar: and the work we are doing so
77 00:13:57.110 ⇒ 00:14:00.690 Awaish Kumar: overall, it’s good. But we are still having some
78 00:14:01.720 ⇒ 00:14:04.520 Awaish Kumar: investigation questions. We want to like.
79 00:14:05.440 ⇒ 00:14:09.089 Awaish Kumar: create a knowledge base where we can put all this
80 00:14:09.842 ⇒ 00:14:13.140 Awaish Kumar: together. So like, we can answer questions.
81 00:14:13.822 ⇒ 00:14:17.219 Awaish Kumar: Timely. But yeah, everything going good
82 00:14:18.660 ⇒ 00:14:25.339 Aakash Tandel: Yeah, I think I’ll add that Josh said this was the best work week with us that we’ve had so far. So that’s good. Definitely moving the ball
83 00:14:25.340 ⇒ 00:14:26.010 Uttam Kumaran: Wow!
84 00:14:26.440 ⇒ 00:14:44.550 Aakash Tandel: Yeah. He’s been joining a lot of the stand ups they do have a like a I think it’s either quarterly call that like 2 o’clock to go over their numbers and their tableau dashboard. So, hoping everything goes well there, and we have no surprises. But yeah, everything else. I wish there was like spot on
85 00:14:45.700 ⇒ 00:14:48.580 Uttam Kumaran: Nice job like big, big big win.
86 00:14:51.053 ⇒ 00:14:53.879 Uttam Kumaran: Okay for Javi. Maybe I’ll pick on
87 00:14:54.050 ⇒ 00:14:58.660 Uttam Kumaran: Kai, are you on? Do you want to give an update on, on how Javi is going
88 00:15:00.180 ⇒ 00:15:01.590 Caio Velasco: Yes, same one
89 00:15:02.670 ⇒ 00:15:13.960 Caio Velasco: So I’ve been noticing that Aman is quite active in the chat, at least for next week, which I think is good and we have been trying to do some
90 00:15:13.970 ⇒ 00:15:33.880 Caio Velasco: some other things regarding the sources we have, and and they have been responsive. But they haven’t done much progress on that part. So we’re still waiting for them. And then I’m gonna also put another person that I didn’t know so to to help us, so it seems to be going well, we we had some progress also this week.
91 00:15:34.415 ⇒ 00:15:39.720 Caio Velasco: at least on on the engineering side, but I also saw that any, and Robert did quite a few things
92 00:15:39.880 ⇒ 00:15:44.349 Caio Velasco: as well. So for me, I had a good feeling this week.
93 00:15:46.020 ⇒ 00:16:12.139 Uttam Kumaran: I agree, I think. I have a very similar feeling to to Eden, and that we’re picking things up. I think my one piece of feedback on Javi is they are. They seem to be the team with like quite a lot of requests. Try this tool, make this change. Everything needs to go through some sort of triage process and get into a pipeline and ideally, you know, we need to turn. We need to basically be able to show them. Hey, you’re requesting
94 00:16:12.480 ⇒ 00:16:29.739 Uttam Kumaran: 10 or 15 new tickets a week we are able to execute on 50% of those with our capacity. If you want us to move faster, you need to pay us more. That’s the perfect step forward here. In the past. We’ve just taken on the work and sort of just figured it out. But.
95 00:16:30.100 ⇒ 00:16:50.859 Uttam Kumaran: you guys, everybody who’s on engineering knows that it’s painful to to get all those things and be like. We can only move as fast as we can. But this is where I think the 3 legged stool is working where we have everything ticketed. We are executing at a good pace. Now I think my pushback is for sales to say cool where we need to have a conversation with how we can
96 00:16:51.570 ⇒ 00:16:52.730 Uttam Kumaran: double our budget.
97 00:16:53.177 ⇒ 00:17:10.689 Uttam Kumaran: So I think that is the conversation we want to be having with all of our clients, and that they want us so much they want. They wanted more from us. They hired this another person, that person is actually gonna show them how difficult some of this work is. And then, you know again, ideally, we we expand our relationship with them.
98 00:17:11.457 ⇒ 00:17:14.579 Uttam Kumaran: Really, really awesome thanks for the update on
99 00:17:14.589 ⇒ 00:17:14.979 Caio Velasco: Sure.
100 00:17:14.980 ⇒ 00:17:35.413 Uttam Kumaran: Pool parts. I think we got a we got a little bit of an update, I think, though, the one piece that I’ll share is across all of our clients, as I mentioned and maybe I I should add this to one of our slides every week, just to talk about how we’re orchestrating the client pods, again, is we have the product owner. We have a project manager. We have engineers.
101 00:17:35.910 ⇒ 00:17:58.290 Uttam Kumaran: it really requires these 3 pieces to to function. A great, a great client pod. I think we on Eden Josh is there now on urban stems. We have that, I think, on pool parts. We were lacking that for a long time, and I was really acting as that Bridge Pool parts is our oldest client. I started the company based on money coming in from them.
102 00:17:58.450 ⇒ 00:17:59.380 Uttam Kumaran: But
103 00:17:59.540 ⇒ 00:18:20.190 Uttam Kumaran: I never even had any idea about this orchestration in 2023 when we decided to work with them. So we, I think today we had a great success, and that Kim, from their side, who runs marketing, has decided to join us in our stand ups, and I do expect this to go smoother. So very, very excited. There. Abc, home.
104 00:18:20.420 ⇒ 00:18:23.879 Uttam Kumaran: Casey, are you on? Do you want to give a little bit of an update
105 00:18:26.409 ⇒ 00:18:32.359 Casie Aviles: Yeah, sure. So I think, yeah, they’re good. I wasn’t on the meeting with them earlier. But
106 00:18:32.740 ⇒ 00:18:35.090 Casie Aviles: yeah, we’re able to focus on
107 00:18:35.520 ⇒ 00:18:41.675 Casie Aviles: like triaging there. You know, when error comes, errors comes in. Comes in and
108 00:18:42.780 ⇒ 00:18:46.611 Casie Aviles: yeah, I guess we also managed to lower some of the
109 00:18:47.230 ⇒ 00:18:53.280 Casie Aviles: error rates. And yeah, I think it’s yeah. We’re just continually improving the bot
110 00:18:55.150 ⇒ 00:19:01.100 Uttam Kumaran: Great and then on Stack Blitz. So one of the I think that
111 00:19:01.360 ⇒ 00:19:11.960 Uttam Kumaran: we’re in sort of a transition period. They hire 2 more data people on their side who who are pretty good. I think we’re gonna see sort of how our relationship shakes up.
112 00:19:12.470 ⇒ 00:19:15.090 Uttam Kumaran: Commonly, it’s it’s like.
113 00:19:15.500 ⇒ 00:19:20.339 Uttam Kumaran: we’re sort of going to see whether they want to boot us, or they want to sort of keep us for some side stuff.
114 00:19:20.470 ⇒ 00:19:33.139 Uttam Kumaran: This one. I’m not sure they’re really fast growing company. So I have a feeling that they’re gonna still need us. Also, you know, I have a really good relationship with the leader here. So let’s see, I’m excited for this to continue to grow.
115 00:19:33.690 ⇒ 00:19:45.049 Uttam Kumaran: I wanted to share just one one thing before I hop off before. I just move to the next slide. And give me a second
116 00:19:46.990 ⇒ 00:19:51.099 Uttam Kumaran: I wanted to share these.
117 00:19:51.590 ⇒ 00:19:55.850 Uttam Kumaran: Where did I put the actual like architecture diagrams.
118 00:19:56.806 ⇒ 00:20:00.190 Uttam Kumaran: Maybe in in data team.
119 00:20:03.520 ⇒ 00:20:10.580 Uttam Kumaran: I have no idea. Does anyone know, I I tossed in those like architecture diagrams that the marketing team developed?
120 00:20:15.320 ⇒ 00:20:15.880 Hannah Wang: I know there’s
121 00:20:15.880 ⇒ 00:20:16.730 Uttam Kumaran: I just wanted to share
122 00:20:16.730 ⇒ 00:20:18.600 Hannah Wang: Panel. Oh, you found it!
123 00:20:18.600 ⇒ 00:20:19.549 Uttam Kumaran: Okay, cool. Alright.
124 00:20:20.240 ⇒ 00:20:35.730 Uttam Kumaran: So I just wanted to share this, and especially for the engineering team. The marketing team is now has a process to take in new requests for architecture diagrams. Execute them, and allow you to then take this and present it to clients.
125 00:20:35.820 ⇒ 00:21:05.029 Uttam Kumaran: Huge, huge win, we all know on the engineering side how hard it is to develop something like this. Not only just develop it, but have it look like, really, really beautiful. This is like what we bring to the table for clients. One, my ask is, if you’re an engineer on on either the job here at Eden, client to please spend a moment just Qa, this, if you have any suggestions for design changes, please let us know. But of course, if we’re going to need a process to take in new requests.
126 00:21:05.721 ⇒ 00:21:11.509 Uttam Kumaran: I think I’ll work with the Pm team to basically think about on a monthly basis. How do we keep this up to date.
127 00:21:11.880 ⇒ 00:21:29.400 Uttam Kumaran: But the other piece I’ll I’ll sort of talk about as I flash up the the Eden version is basically that we’re gonna start to go deeper, meaning. We’re gonna have full architecture diagrams that include every single core marts model
128 00:21:29.850 ⇒ 00:21:52.639 Uttam Kumaran: as well as once we get the bi layer sort of modeled out, we’ll be able to see the core dashboards as well. This is going to be our home for talking about what architecture is for a client. We’re going to be developing this for pool parts for ABC, which is going to be sort of a different set of architecture. And then, additionally, we’re going to do before and afters right and the before and after is really where we’re going to show
129 00:21:52.640 ⇒ 00:22:02.780 Uttam Kumaran: what for urban stems. This is a really great example is, we’re doing it before we’re going to do it after. And then this will sort of set the stage for us to say, here are the number of tools. Here are the costs for all of them.
130 00:22:02.990 ⇒ 00:22:27.939 Uttam Kumaran: and like if I was, this is finally, I think, one of the points where, if I was hiring Brainforge and I got this. I’d be very, very impressed. And so please just take a moment out of your day to Qa. This make sure if you’re an engineer on this client that these match up your expectations. But we’re gonna start having these for every client ideally have a before version, you know. And then, as we work on a client, this will get updated.
131 00:22:30.970 ⇒ 00:22:36.360 Uttam Kumaran: Cool. Great any other questions here.
132 00:22:40.320 ⇒ 00:22:52.727 Uttam Kumaran: Okay, so this is where I kind of wanted to make this a little bit more collaborative. Does anyone else? You can feel free to put it in the chat, but I wanted to sort of talk through accomplishments this week.
133 00:22:53.270 ⇒ 00:23:20.835 Uttam Kumaran: finally, this week is the week where I really feel like I I wasn’t able to keep up with every team, which is a good thing. Like I. And you know it’s actually the best thing is I get to read all the meeting notes that come in from the zoom. Bot so I’m very happy. But I need some help sort of putting this together, and I probably will start to get crowd sources a bit at the end of every week. But does anyone have any other accomplishments? We want to
134 00:23:21.230 ⇒ 00:23:24.800 Uttam Kumaran: point in here beyond what I already wrote.
135 00:23:25.880 ⇒ 00:23:34.859 Uttam Kumaran: and this is a good great log for us historically, to sort of keep track. But if you want to just write in the chat or say it out loud, I will. I’ll update it here.
136 00:23:41.810 ⇒ 00:23:42.940 Uttam Kumaran: Yes.
137 00:24:06.530 ⇒ 00:24:11.459 Uttam Kumaran: yes, we did. So yeah, we have a new accounting team that’s coming on as well
138 00:24:13.910 ⇒ 00:24:16.590 Nicolas Sucari: I don’t know if it is an accomplishment, but
139 00:24:16.590 ⇒ 00:24:19.220 Uttam Kumaran: It is. Yeah, I mean, it’s like I,
140 00:24:19.500 ⇒ 00:24:31.970 Uttam Kumaran: you’d be surprised that it takes a long time to figure that out. And that’s not my core competency. My core competency is data. So we want to have a great team that’s handling all of our accounting, finance, bookkeeping.
141 00:24:36.000 ⇒ 00:24:40.090 Uttam Kumaran: Any other accomplishments.
142 00:24:49.190 ⇒ 00:25:18.470 Uttam Kumaran: I I really do think that we’re using linear a lot better. I think my only suggestion and this is always gonna be a conflict is for engineers to really push and and have great requirements in your tickets before you start working on them as well, has have set due dates and set estimations. This is gonna really help understand? Like, how much capacity we can take on but I’m seeing linear being used across all teams, including operations, marketing, recruiting.
143 00:25:29.650 ⇒ 00:25:38.079 Uttam Kumaran: Okay, great. And then maybe we could talk about challenges. Anyone. Wanna throw out any like big
144 00:25:38.330 ⇒ 00:25:44.670 Uttam Kumaran: challenges that you face this week can be personal can be about your team or about the company.
145 00:25:51.270 ⇒ 00:26:07.970 Uttam Kumaran: for me. One of the pieces that I really learned is that of course not. Everyone is familiar with our core stack. And so I think we need to have some level of continuous learning, like a continuous learning program where folks can get certified on every tool, and I
146 00:26:08.230 ⇒ 00:26:30.309 Uttam Kumaran: I’m not certified. I I was looker certified at some point. I’m not sure if it that elapsed, but it’d be great for all of us to to get certified on our core tools. Happy to pay for any certification. Exam there. And then also sort of building a channel where we can all can. All ask questions. Yeah, definitely have it as part of onboarding
147 00:26:31.560 ⇒ 00:26:34.279 Uttam Kumaran: any other challenges from anyone this week
148 00:26:45.292 ⇒ 00:26:50.079 Miguel de Veyra: I think one thing is keeping you on the loop with them about the progress of the AI team.
149 00:26:50.990 ⇒ 00:26:53.019 Miguel de Veyra: Yeah? Cause? Yeah, yeah.
150 00:26:57.070 ⇒ 00:26:58.799 Nicolas Sucari: And having maybe the
151 00:26:59.020 ⇒ 00:27:06.619 Nicolas Sucari: habit of going into linear every day to see progress and move stuff there, so that we can show that to everyone
152 00:27:08.640 ⇒ 00:27:11.857 Uttam Kumaran: Yeah, maybe I’ll just speak. Speak about this
153 00:27:13.290 ⇒ 00:27:29.179 Uttam Kumaran: like this week was again the the 1st week where I really felt that I had to rely on linear and our Zoom Meeting summary bots to understand how things are going. And A lot of probably the challenges were like, Hey, where’s progress here
154 00:27:29.260 ⇒ 00:27:56.509 Uttam Kumaran: for me? I think the reason why we use these systems is not for micromanagement, but actually for us to scale our understanding of progress, meaning. Everybody has the same definition of progress. What success is and it allows us to go from our 6 or 7 clients to 17 clients to 27 clients, and for some for me to find out where I need to spend my time. Right? There are clients now that are going well where I’m comfortable.
155 00:27:56.610 ⇒ 00:28:18.169 Uttam Kumaran: saying just, I’ll read the updates at the end of the day. That is not how at all how we’ve run the company in the past, and it’s really great for me, because then I can spend my time on things that are higher leverage, which is spending time on continue learning hosting office hours, where we can focus on key challenges, bringing in external support and coaching
156 00:28:18.505 ⇒ 00:28:30.400 Uttam Kumaran: and then going and getting us clients. Those are the highest leverage activities for Robert and I. And it’s really important that we have these tools up to date, you know, and we can scale these systems.
157 00:28:30.560 ⇒ 00:28:36.419 Uttam Kumaran: This is otherwise. If you guys are worked at sort of companies where you don’t have these systems.
158 00:28:36.880 ⇒ 00:28:42.550 Uttam Kumaran: what happens is you spend more time in meetings having to discuss updates
159 00:28:43.270 ⇒ 00:28:55.960 Uttam Kumaran: versus actually working. And that’s 1 thing that I want to really avoid is like status update meetings. In fact, I expect our linear usage to get so good that we could probably start to skip some standups
160 00:28:56.278 ⇒ 00:29:18.799 Uttam Kumaran: and start to space out meetings more which gives people a lot more time back. And we can have meetings that specifically about a problem where we all need to be together. That’s the ultimate Holy Grail, right for engineering is that the tickets are everybody’s updates are there so that everybody knows what they need to do next. There’s nothing to talk about right. And then we talk more about challenges in the future. And so this is really really important.
161 00:29:21.100 ⇒ 00:29:25.960 Uttam Kumaran: Any other challenges this week it could be small. It could be big. Anything
162 00:29:33.910 ⇒ 00:29:36.469 Caio Velasco: Oh, I can share that on my end.
163 00:29:37.580 ⇒ 00:30:06.019 Caio Velasco: it’s always gonna be a challenge, even if it sometimes it’s just a small thing to understand. Like some of the sources. For example, I see that Amazon has lots of documentation online. But still you have to like, really spend a lot of time going to the developer section and trying to see how those things are connect to the tables and and the sources we have. And then, at the end of the day you end up losing time you learn a lot which I think long term will pay off.
164 00:30:06.070 ⇒ 00:30:08.820 Caio Velasco: But yeah, it’s always a challenge for me.
165 00:30:10.350 ⇒ 00:30:31.010 Uttam Kumaran: Yeah. So this is something that I’m going to push as a really great opportunity for the AI team, and probably something that we’ll spend the rest of this call. Talking about is like where we’re going to try to apply. AI. Part of this is going to be. My expectation is that we’re going to start to build agents that have ingested this documentation for our core sources that we can use to ask.
166 00:30:31.010 ⇒ 00:30:44.990 Uttam Kumaran: Where should I go to learn this. And what are the answers for this? I think, to really break this down. If Kyle has spent, you know, 2, 3 h, and I know if you go to. This has happened all of our sources where you have to go, really, really in depth.
167 00:30:45.210 ⇒ 00:30:57.030 Uttam Kumaran: and we’re able to have an AI agent that ingests all that docs and sort of can answer that question in just a few seconds. That’s a huge, huge win. And so that probably leads me to my
168 00:30:57.170 ⇒ 00:31:05.961 Uttam Kumaran: my next sort of section. Here is where I want to. I want to share, you know, probably 2 2 things that were really critical.
169 00:31:07.270 ⇒ 00:31:23.430 Uttam Kumaran: and were really important for me to to see this week from the AI team, which was, which is basically our AI team goals and and roadmap for Q. 2. And this is where again I want. I want this to
170 00:31:24.160 ⇒ 00:31:42.159 Uttam Kumaran: the way I think about our AI team, and sort of the reason why we formed the team is that we wanted this to be an asset for the company. Right? I think we. I feel very lucky that we’re part of an organization while AI is available. And we can really help us focus on the things that really
171 00:31:42.200 ⇒ 00:32:09.359 Uttam Kumaran: need our critical thinking in our human brain. As you guys know, like 20 to 40% of our tasks are just moving information around updating linear tickets. Basic Google searches. And this is a situation where I want us to try to lean on AI more. I really really encourage everybody to take a look at this document. I will send it here in the chat. Please give this a look when you have a moment.
172 00:32:09.860 ⇒ 00:32:13.929 Uttam Kumaran: I think, above a lot of our efforts.
173 00:32:14.281 ⇒ 00:32:43.330 Uttam Kumaran: The effort that we’re going to put in. And the investment we’re going to put in to make AI a core resource for all of our team members is going to be an extremely extremely highly leveraged opportunity for us. And it’s going to really differentiate us, not only as a great place to work where we can actually focus on pushing great code, but also our ability to scale pretty quickly and to start to be more of a value. Add to our clients
174 00:32:43.530 ⇒ 00:32:56.260 Uttam Kumaran: one of the things that I really wanted to share is sort of the key focus areas for the AI team. And in particular, I wanted to talk about the internal AI clients.
175 00:32:56.520 ⇒ 00:33:24.339 Uttam Kumaran: The narrative here is that the AI team is an internal organization and its clients are other teams. This builds this. This allows us to sort of basically build an internal like, basically an internal resource that can be tapped in from multiple teams when you need to use AI in your day to day. Additionally, this team, their sort of goal is to go to every team and say, How can we solve problems for you?
176 00:33:24.630 ⇒ 00:33:46.679 Uttam Kumaran: The goal of the team, you know, that’s that’s having problems being solved is just to narrate how you’re spending your day like, where’s where’s a lot of time going that you wish you had back common problems you face every day. And the AI team job is to find out, okay, where are there opportunities for automation here? And really, I think you know, Amber Miguel Casey
177 00:33:47.000 ⇒ 00:34:02.910 Uttam Kumaran: did an extremely good job sort of articulating here. How we’re going to focus on this in the next quarter, really measuring that the agents, the automations that we develop for each team are getting used, but also that they’re they’re showing a clear roi
178 00:34:03.060 ⇒ 00:34:18.305 Uttam Kumaran: from my perspective. I want us to see that by us being able to work on more clients, with the same amount of people. But also for us to actually go deeper on like core work, right? So shaving off the hours that we’re spending
179 00:34:18.790 ⇒ 00:34:24.039 Uttam Kumaran: like just transferring information between sources. And I know it’s a little bit high level
180 00:34:24.600 ⇒ 00:34:26.229 Uttam Kumaran: but I wanted to share this
181 00:34:26.704 ⇒ 00:34:35.530 Uttam Kumaran: with everyone that we’re be. We’re going to be focused as an AI team on affecting the sales and data teams. First, st we will then sort of move pretty
182 00:34:35.699 ⇒ 00:34:43.140 Uttam Kumaran: a regimently across operations, marketing finance. And so I’m I’m very, very excited for this.
183 00:34:43.830 ⇒ 00:34:47.779 Uttam Kumaran: Any questions here before I just talk about the next section.
184 00:34:50.150 ⇒ 00:35:01.159 Uttam Kumaran: This is a very, very new like this. Technology has only been available for about 2 years. So it’s extremely new. So please, if you have any questions about this.
185 00:35:03.767 ⇒ 00:35:05.110 Uttam Kumaran: could ask away.
186 00:35:11.290 ⇒ 00:35:12.160 Uttam Kumaran: Okay.
187 00:35:12.530 ⇒ 00:35:26.739 Uttam Kumaran: the next piece. And this is actually something that I thought a lot about this week is actually beyond just the AI team building things is actually just educating everybody on how they can use AI in their day to day. And I actually feel like
188 00:35:27.040 ⇒ 00:35:34.410 Uttam Kumaran: this is probably equally an equivalent amount of impact to just the AI team building stuff, which is.
189 00:35:34.430 ⇒ 00:36:04.157 Uttam Kumaran: can we start to do workshops on, how to use cursor, how to use the zoom, the Zoom transcript agent. You know. How do we? How do the best people in the company. The most efficient people use chat. Gpt, so I spend a lot of my time in these tools, and and I want everybody to sort of be educated with, how how they can get their jobs done faster, and how they can spend more time doing things that are actually fun and actually interesting for all of us, you know, and and spending time avoiding the things that are boring or easy to solve.
190 00:36:04.900 ⇒ 00:36:23.519 Uttam Kumaran: you know. And for me. I want to reiterate this. I don’t see AI as like, Okay, we’re going to replace everybody. That’s like a kind of a tired narrative. I think it’s kind of like a foolish errand. It’s like not the point at all. The point for me is that this is not not. This is not a binary equation like it’s not
191 00:36:23.880 ⇒ 00:36:33.179 Uttam Kumaran: have everybody or replace everybody. What I think a lot about is taking care of 20% of the really boring, annoying
192 00:36:33.310 ⇒ 00:36:37.159 Uttam Kumaran: sort of like low, level information transfer work
193 00:36:37.290 ⇒ 00:36:45.400 Uttam Kumaran: that we can lean on AI for and helping us spend our time on working on really tough problems or with the client.
194 00:36:45.520 ⇒ 00:36:48.128 Uttam Kumaran: That’s it, right? And so,
195 00:36:48.790 ⇒ 00:36:55.319 Uttam Kumaran: we’re working on, you know, creating a couple of initiatives here. But does anyone have any questions on this?
196 00:36:55.480 ⇒ 00:36:56.660 Uttam Kumaran: At the moment.
197 00:37:06.330 ⇒ 00:37:11.628 Uttam Kumaran: Great. I think one one piece I want to. Just
198 00:37:12.480 ⇒ 00:37:36.476 Uttam Kumaran: I want to just share is is this AI roadmap that? I know the data team, a few folks and the AI team worked on. I mean, this may be, you know, pretty this. I would encourage everybody on the data team to read, and we will. This 1st sort of set of feedback. All of this will get exposed to everybody, and we’ll ask everybody for feedback. Of course.
199 00:37:37.390 ⇒ 00:37:52.330 Uttam Kumaran: what I, what I task the AI team to do is basically go interview and sort of understand what the core challenges of the data team are, and sort of understand where we can, we can affect that. And so we do have a lot of really great suggestions here, which is
200 00:37:53.225 ⇒ 00:38:03.809 Uttam Kumaran: around documenting dpt models, checklist generator for easy tasks. Query assistant Qa. Agents.
201 00:38:04.300 ⇒ 00:38:18.360 Uttam Kumaran: This, doc, I literally was like smiling, and then also like crying, reading this document, because this was like, if we did all these things it would be insane like, we’re staring at a document right? None of this is real yet, but
202 00:38:18.830 ⇒ 00:38:21.370 Uttam Kumaran: I think for everybody in the AI team.
203 00:38:21.870 ⇒ 00:38:45.015 Uttam Kumaran: The reason why these are medium and and quick is that it’s just the time to do it. It’s not the fact that we can’t do any of these things right. Nothing here is like we need to change physics in order to make this happen. And that’s like such a joy to me, because we on the data side, we all know that, like these, these tasks take up so much of our day.
204 00:38:45.710 ⇒ 00:38:47.960 Uttam Kumaran: and I really wanna.
205 00:38:48.480 ⇒ 00:38:57.209 Uttam Kumaran: I really want to put my weight behind making some of these things real and making this like again. I think a lot about, what is it like to be a data engineer
206 00:38:57.320 ⇒ 00:39:02.429 Uttam Kumaran: at Brainforge and having access to some of these, I think, would be really really quite amazing.
207 00:39:05.880 ⇒ 00:39:09.950 Uttam Kumaran: Any questions or thoughts from anyone on the data team or the AI team?
208 00:39:18.500 ⇒ 00:39:23.790 Uttam Kumaran: Okay? Great, I think. What? What we’ll also try to do is oh, yeah, demo on it. Go ahead.
209 00:39:24.770 ⇒ 00:39:30.839 Demilade Agboola: I was just gonna say that this is, I mean, obviously still a work in progress. But, like
210 00:39:31.613 ⇒ 00:39:39.860 Demilade Agboola: The AI team, especially like Amber, was able to articulate like the problem statements and figure out ways to be able to apply
211 00:39:40.080 ⇒ 00:39:46.010 Demilade Agboola: AI to some of the common problems that we have so obviously
212 00:39:46.330 ⇒ 00:39:56.240 Demilade Agboola: as we expose it to the rest of the team. If you have any other problems or things that hold up your work a lot more than these. Please feel free to also mention
213 00:39:59.400 ⇒ 00:40:07.259 Uttam Kumaran: Yeah, I think that’s a great point, you know, I think about back to when I was doing, you know, more like product management stuff is the the goal here is 1st to
214 00:40:07.260 ⇒ 00:40:29.850 Uttam Kumaran: get it, like the 80% of like what the core problems are and how we’re going to solve it. The next piece is, we’re gonna circulate this to everybody, right? So in one meeting or another, this document is going to be there, and you’re going to be asked for comment. I really really hope that everybody on the data side can take can just take 5 min and look at this and make sure that your concerns are are in here, and that if we do solve one of these
215 00:40:29.850 ⇒ 00:40:34.950 Uttam Kumaran: it will have an impact on your day. And so we’ll make sure to circulate this as well.
216 00:40:35.180 ⇒ 00:40:43.109 Uttam Kumaran: Maybe I know we just have a little bit of time left. I was gonna pick on Amber or Miguel just to share a little bit about
217 00:40:43.230 ⇒ 00:40:44.040 Uttam Kumaran: like
218 00:40:44.160 ⇒ 00:41:01.180 Uttam Kumaran: a couple of ways you’re using AI in your day to day. I know we talked about Granola. We talked about cursor. We talked about just the Chat Gpt ui. But I sort of want to carve out a piece of every week now, just to show how people are using AI in interesting ways.
219 00:41:01.410 ⇒ 00:41:08.399 Uttam Kumaran: And yeah, maybe I’ll just like, leave the floor open for both of those folks. And then anyone else that wants to sort of
220 00:41:09.011 ⇒ 00:41:14.789 Uttam Kumaran: just share like cool ways. You’re using AI, it can be personally, it can be professionally. But yeah.
221 00:41:16.960 ⇒ 00:41:31.040 Amber Lin: I can get started. So I’ll talk about Granola, and I know most people use chat. Gbt, I’ll talk about Granola, and I’ll talk about internal agents, because I will share my screen here.
222 00:41:31.600 ⇒ 00:41:40.169 Amber Lin: We luckily, I’m on the ABC. Team, and we make the agent for ourselves because we can. So here.
223 00:41:40.330 ⇒ 00:41:47.409 Amber Lin: I asked it a few things one I wanted to onboard me because we’re we were onboarding Annie this week.
224 00:41:47.870 ⇒ 00:41:51.897 Amber Lin: and I asked them, Hey, we I have
225 00:41:52.830 ⇒ 00:42:00.120 Amber Lin: what? What’s going on? Right? So I gave it a list of questions, a long list of questions.
226 00:42:00.270 ⇒ 00:42:05.569 Amber Lin: and it just answered everything. So is that org chart
227 00:42:06.040 ⇒ 00:42:09.759 Amber Lin: past projects. What tools do we use?
228 00:42:09.910 ⇒ 00:42:16.999 Amber Lin: Current projects? The time on budget? This is so so detailed because it has context to
229 00:42:18.120 ⇒ 00:42:23.639 Amber Lin: everything that went on in slack and zoom. So this saved me so much time
230 00:42:24.650 ⇒ 00:42:28.419 Uttam Kumaran: And amber. Can you talk about how you’re actually accessing this in slack
231 00:42:29.224 ⇒ 00:42:45.759 Amber Lin: So right now, this is in AI agents. I am aware that not everybody is in there. So in slack, you can add, you can add this agent to your channel, so say if I click here.
232 00:42:46.230 ⇒ 00:43:08.010 Amber Lin: I guess you can add this app to a different channel. I’ll let Casey, mango, explain this later, and we’ll send out like a tutorial on how to use these things. Essentially. You add it in a message, and you say you, you prompt it essentially takes a bit of prompting to tell them what who you are and what you’re looking for.
233 00:43:08.560 ⇒ 00:43:13.320 Amber Lin: So in the next part. I also asked it. I need to make a slide deck.
234 00:43:13.810 ⇒ 00:43:16.679 Amber Lin: and I told them things to consider.
235 00:43:17.480 ⇒ 00:43:25.229 Amber Lin: And then it helped me create this. And this is essentially
236 00:43:25.790 ⇒ 00:43:34.079 Amber Lin: the core thing that I used to create the presentation and put it through chatgy routine, and I fine tune it a little bit. But it was really really helpful.
237 00:43:34.790 ⇒ 00:43:45.979 Amber Lin: and I talked to Robert. And we’re like, Oh, we don’t have that much case studies. And this is a case study that I made it right, so gave it a structure that I copy and paste from a Google search.
238 00:43:46.170 ⇒ 00:43:53.689 Amber Lin: And then it gave a pretty nice answer of, Okay, what are the results.
239 00:43:53.930 ⇒ 00:43:57.719 Amber Lin: What are the methodologies objectives?
240 00:43:58.160 ⇒ 00:43:59.150 Amber Lin: So
241 00:43:59.460 ⇒ 00:44:08.583 Amber Lin: because they have so much, these agents have so much context of what we’re doing, it really helps speed up a lot of
242 00:44:09.790 ⇒ 00:44:29.309 Amber Lin: things related to writing documentation. And our goal for the AI team is we want to do that for your data models as well, because when we talk to the data team, big problem is we don’t have documentation to send to the client or to give to new analysts. And most of your time is spending
243 00:44:29.560 ⇒ 00:44:42.490 Amber Lin: spent on something that’s already there in the system. But you have to figure it out. So when that’s 1 of our goals to just give you guys the context and give you guys the tools to make that happen really fast.
244 00:44:42.610 ⇒ 00:44:47.489 Amber Lin: So that’s on my end. I’ll let Miguel share a bit on how he uses AI
245 00:44:48.060 ⇒ 00:44:51.889 Miguel de Veyra: Hey guys? Yeah. So the way I use it. Because
246 00:44:52.830 ⇒ 00:45:14.859 Miguel de Veyra: actually, recently, I’ve been like, not really on development, like, I’ve been on writing like technical requirements and stuff. So I I primarily use Chat Gpt, for now and then I created like separate projects, because it’s also my job to do recruiting for the AI team now and then. Basically, I created, let me just share my screen actually
247 00:45:19.500 ⇒ 00:45:21.630 Miguel de Veyra: here, and go.
248 00:45:22.380 ⇒ 00:45:28.590 Miguel de Veyra: So basically, I create, I just created separate projects and then be
249 00:45:29.160 ⇒ 00:45:47.559 Miguel de Veyra: right. Now, I’m still exploring the address, I mean for recruitment. I don’t really need anything there. But basically, this is just a quick fire away to create your own AI agent. So I created one for recruitment, right? And then, basically, I gave it like the process I have on recruiting. So basically, I just gave it this.
250 00:45:48.030 ⇒ 00:45:51.930 Miguel de Veyra: So it has that context and then perfect.
251 00:45:52.480 ⇒ 00:45:57.219 Miguel de Veyra: And then you know what we’re looking for. This is basically the Jd that I worked on
252 00:45:57.470 ⇒ 00:46:03.270 Miguel de Veyra: so from there on, like, I think this is for Vasu. I’ve already contacted him.
253 00:46:03.390 ⇒ 00:46:11.699 Miguel de Veyra: and then the other. There’s I think that one more guy, and then for this one is, it’s a bit more for the there you go
254 00:46:12.120 ⇒ 00:46:15.340 Uttam Kumaran: Wait, Miguel, can you slow down like a hundred percent
255 00:46:15.874 ⇒ 00:46:16.690 Miguel de Veyra: Okay, yeah. My, pat.
256 00:46:16.690 ⇒ 00:46:18.750 Uttam Kumaran: Like a hundred like 200%.
257 00:46:19.261 ⇒ 00:46:24.239 Uttam Kumaran: Go, wait, go start, start by just explaining like what you were trying to do.
258 00:46:24.390 ⇒ 00:46:25.359 Miguel de Veyra: Oh, yeah, I think there’s
259 00:46:25.360 ⇒ 00:46:34.060 Uttam Kumaran: A lot of people even beyond, just like just using Chat Gpt, but but also like building agents and projects. So just start with the goal
260 00:46:34.510 ⇒ 00:46:38.020 Miguel de Veyra: So my goal was, that’s I think the recruitment would be a lot easier to explain.
261 00:46:38.180 ⇒ 00:46:39.939 Miguel de Veyra: Like for recruitment.
262 00:46:39.940 ⇒ 00:46:42.240 Uttam Kumaran: And can you zoom in? Can you zoom in just a little bit
263 00:46:42.240 ⇒ 00:46:43.369 Miguel de Veyra: Yeah, not bad.
264 00:46:43.610 ⇒ 00:46:44.720 Miguel de Veyra: Is this better?
265 00:46:44.720 ⇒ 00:46:45.630 Uttam Kumaran: Yeah, yeah.
266 00:46:45.630 ⇒ 00:47:02.310 Miguel de Veyra: Okay. So for recruitment, for example, instead of like, you know, if I create like a new one, I have to re prompt, basically provide context like, Hey, here’s the let’s go here that here’s all the stuff that I have. Here’s the stages right? Here’s the person. Here’s the you know, whatever.
267 00:47:02.970 ⇒ 00:47:05.860 Miguel de Veyra: So what I did was I created like a project?
268 00:47:06.561 ⇒ 00:47:13.619 Miguel de Veyra: So you can just do that to here new project Yada Yada, name it. And then it’s gonna bring you here. It has 2 things. This
269 00:47:13.810 ⇒ 00:47:19.270 Miguel de Veyra: I’m not sure if everyone is familiar with knowledge bases. So basically, this is how you add it here
270 00:47:20.680 ⇒ 00:47:27.170 Miguel de Veyra: and then instructions, basically, system prompt where it’s the Pre, the preset instructions.
271 00:47:27.560 ⇒ 00:47:30.659 Miguel de Veyra: And then you just provided you know
272 00:47:30.970 ⇒ 00:47:49.279 Miguel de Veyra: a set of instructions. So you don’t have to retype this every time that you have to. For example. Me, I’m reaching out to 1, 2, I would say, 3, 4, yeah, there’s a lot of people I’m reaching out to right now. So I don’t want to retype this or copy paste this from somewhere. Right? So now, if I just, for example.
273 00:47:49.900 ⇒ 00:47:51.669 Miguel de Veyra: type a name here so
274 00:47:52.510 ⇒ 00:48:07.454 Miguel de Veyra: super fast. Why, just basically paste the emails here and everything, and then try to, of course, prompt it around, play around with a bit. It’s a lot easier. It’s a lot, you know, straightforward. There’s actually a structure already. So that’s 1 of the ways I use. AI
275 00:48:08.510 ⇒ 00:48:12.649 Miguel de Veyra: on my day to day. But yeah, this was probably
276 00:48:12.780 ⇒ 00:48:15.760 Miguel de Veyra: the one I use a lot more, as you can see.
277 00:48:17.274 ⇒ 00:48:23.079 Miguel de Veyra: This basically just helps me. I provided again instructions same as the one before.
278 00:48:23.860 ⇒ 00:48:26.970 Miguel de Veyra: And then I provided it like a good.
279 00:48:27.980 ⇒ 00:48:30.380 Miguel de Veyra: basically technical. How do I go?
280 00:48:31.480 ⇒ 00:48:33.210 Miguel de Veyra: Technical requirements? So
281 00:48:34.230 ⇒ 00:48:59.480 Nicolas Sucari: Question, there, Miguel, real quick. So creating a project is like having your own agent with different contexts. Right? Like you kind of give different context to each of the projects, so that it can help, like solve specific of specific things. On each of those projects you have recruiting with all the context regarding recruiting files. That you’re looking for. And then you have, like your internal one, with different instructions or context, to solve different things right
282 00:48:59.480 ⇒ 00:49:00.520 Miguel de Veyra: Yes, yes.
283 00:49:00.740 ⇒ 00:49:07.639 Miguel de Veyra: So basically, you know, instead of ha me having to go to any 10 and build this one because it’s pretty much the same.
284 00:49:09.320 ⇒ 00:49:12.120 Miguel de Veyra: you know, like even here, you can connect it.
285 00:49:13.030 ⇒ 00:49:25.930 Miguel de Veyra: It’s just, you know, it’s faster to do it here, especially. There’s already a ui right, and I don’t have to connect it to either slack, to Google chat or to a Ui that we need to build. So yeah.
286 00:49:27.100 ⇒ 00:49:31.800 Miguel de Veyra: I guess this is this is something we should probably look into.
287 00:49:32.885 ⇒ 00:49:34.660 Miguel de Veyra: Implementing for
288 00:49:34.940 ⇒ 00:49:38.550 Miguel de Veyra: one of the team one of the teams we wanna try. Maybe ABC is a good start
289 00:49:38.710 ⇒ 00:49:40.330 Miguel de Veyra: just to test it out. But yeah.
290 00:49:41.770 ⇒ 00:49:44.010 Miguel de Veyra: I should probably create an ABC project here
291 00:49:44.740 ⇒ 00:49:48.360 Nicolas Sucari: Does this require to have a paid plan of chatgpt
292 00:49:48.690 ⇒ 00:50:00.449 Uttam Kumaran: Yeah. So what? I’m so so I think short term, I’m happy to pay for anyone who wants to just have chat, gpt pro so just ping me, actually, yeah, just ping me and Nico, we can have the
293 00:50:00.450 ⇒ 00:50:02.660 Nicolas Sucari: Don’t be. Don’t, don’t. Okay. Yeah.
294 00:50:02.680 ⇒ 00:50:04.140 Uttam Kumaran: Okay, go ahead. Yeah. Go ahead.
295 00:50:04.140 ⇒ 00:50:04.530 Nicolas Sucari: Totally.
296 00:50:04.530 ⇒ 00:50:10.336 Uttam Kumaran: Use the use the tool request linear. Ask, do not ping me
297 00:50:11.369 ⇒ 00:50:22.440 Uttam Kumaran: so that’s 1 thing. The second piece is, I actually maybe for Miguel Amber and Casey I actually want to just expose, like Chat gpt via an agent in slack
298 00:50:22.920 ⇒ 00:50:40.449 Uttam Kumaran: like I actually think we can do. Take this one step further where we can just say, Hey, you can just ping like chat, gpt bot! DM it directly or in a channel, and it’ll just act as interface. We’ll just use the Api cause. I think that’s even one step
299 00:50:40.580 ⇒ 00:50:56.050 Uttam Kumaran: further than like people having to open, chat, gpt and then ask a question. If we can just expose just the raw 4 0 Api 4 0 is an Api directly in slack. And just say, this is the Brainforge chat, Gpt, bot.
300 00:50:56.390 ⇒ 00:50:59.500 Uttam Kumaran: use it. And that way we can also probably save.
301 00:50:59.720 ⇒ 00:51:03.379 Uttam Kumaran: It’s like 20 or 30 bucks a month for a person, probably, save that as well
302 00:51:03.380 ⇒ 00:51:05.480 Miguel de Veyra: Yeah, okay, yeah, we can do that
303 00:51:05.480 ⇒ 00:51:06.100 Nicolas Sucari: Yeah.
304 00:51:06.720 ⇒ 00:51:07.260 Uttam Kumaran: Generic.
305 00:51:07.600 ⇒ 00:51:11.239 Miguel de Veyra: Yeah, yeah, it’s a bit more straightforward. I think, yeah, we can definitely do that
306 00:51:11.720 ⇒ 00:51:12.340 Uttam Kumaran: Yeah.
307 00:51:14.622 ⇒ 00:51:20.559 Miguel de Veyra: But yeah, I mean the thing I like here is, it’s organized, right? The conversations are here and everything. But yeah.
308 00:51:23.190 ⇒ 00:51:25.870 Uttam Kumaran: Any any questions from anybody
309 00:51:30.480 ⇒ 00:51:35.240 Luke Daque: I I didn’t even know this. We can do this in chat, gpt like projects.
310 00:51:35.630 ⇒ 00:51:42.259 Luke Daque: I think the only question I have is like, is there any way to like
311 00:51:43.230 ⇒ 00:51:46.160 Luke Daque: like just making the instructions? Basically is like.
312 00:51:47.083 ⇒ 00:51:51.560 Luke Daque: What’s the best approach like, how how do you make good instructions
313 00:51:51.840 ⇒ 00:51:59.580 Miguel de Veyra: Basically just draft something that you would like to tell. Honestly, I just use AI to to improve AI like, it’s
314 00:51:59.580 ⇒ 00:52:01.179 Miguel de Veyra: yeah, like, I don’t know.
315 00:52:01.180 ⇒ 00:52:03.210 Miguel de Veyra: Well, Miguel, I think, for example.
316 00:52:03.410 ⇒ 00:52:07.869 Uttam Kumaran: Yeah, I think one example, Amber. You were even thinking of a prompt library. Basically, right?
317 00:52:09.800 ⇒ 00:52:14.670 Uttam Kumaran: So I think part of that is like there’s some prompts that I use on how to make good prompts.
318 00:52:14.860 ⇒ 00:52:18.340 Uttam Kumaran: So maybe you can talk about like what the prompt library is going to be sorry
319 00:52:19.910 ⇒ 00:52:29.660 Amber Lin: So one of our that we decided on as an easy one is, we’re gonna make a library of prompts that you can just copy and paste
320 00:52:29.760 ⇒ 00:52:46.549 Amber Lin: and ask AI to tailor it for your needs so that you can prompt it better. So we’ll just collect a lot of use cases for different ais. We’ll give you the best structure and then give you like a template to customize that on.
321 00:52:46.970 ⇒ 00:52:55.089 Amber Lin: So if you guys have any suggestions, just feel free to tell us, and we’ll collect that. Maybe we could do a Google form
322 00:52:59.020 ⇒ 00:53:17.979 Miguel de Veyra: But yes, for the question, Luke, just give it like, basically, there’s 3. There’s 3 parts. What’s the goal? What’s the special instructions you want to add, and what’s the output? I’d start there, just put anything you want, and then basically tell Gpt, hey, can you improve the language in this a bit.
323 00:53:18.640 ⇒ 00:53:25.570 Miguel de Veyra: And then and then, you know, say that it’s it’s a prompt basically mentioned that it’s said that the instructions were a bot.
324 00:53:25.750 ⇒ 00:53:31.320 Miguel de Veyra: I find that works a lot of time even for even for ABC, I think that’s our form.
325 00:53:33.500 ⇒ 00:53:36.090 Miguel de Veyra: And that’s that was pretty successful. So yeah.
326 00:53:36.400 ⇒ 00:53:39.389 Luke Daque: Yeah, I wonder if this is something we can do in like
327 00:53:39.620 ⇒ 00:53:45.500 Luke Daque: cursor. Ex example. For, for example, like we make cursor like, always
328 00:53:46.300 ⇒ 00:53:53.419 Luke Daque: think that it’s a data engineer or an analytics engineer like for for my case, for example, because, like, I’m doing like data models or whatnot.
329 00:53:53.660 ⇒ 00:53:55.229 Luke Daque: So that yeah, we don’t.
330 00:53:55.420 ⇒ 00:54:01.799 Luke Daque: It doesn’t like stray away too much from data, engineering and analytics engineering. So yeah, something like that. But yeah.
331 00:54:01.800 ⇒ 00:54:07.810 Miguel de Veyra: You can. You can. Even you can go one step ahead and like behaviors what to do and what not to do.
332 00:54:08.280 ⇒ 00:54:14.609 Miguel de Veyra: and then tell it, basically, hey? Sometimes I do it when there’s too much emojis, you know.
333 00:54:14.980 ⇒ 00:54:17.919 Miguel de Veyra: hey? Don’t add emojis, and then maximum, please.
334 00:54:18.280 ⇒ 00:54:21.140 Miguel de Veyra: for God’s sake, 200 characters you’re over, explained.
335 00:54:22.201 ⇒ 00:54:25.990 Miguel de Veyra: 50% of this I don’t need.
336 00:54:26.440 ⇒ 00:54:31.790 Miguel de Veyra: But yeah, it’s yeah. It’s a, you know. And then I would just use AI to improve the prompt
337 00:54:31.790 ⇒ 00:54:45.399 Amber Lin: Yeah, we’ll we’ll send out. I think this could be something that the AI team does like right right away. We’ll just send out some instructions to educate everybody on how to best prompt, because I always
338 00:54:45.400 ⇒ 00:54:48.220 Miguel de Veyra: Send out my template on the Brainforge team
339 00:54:48.220 ⇒ 00:54:48.610 Amber Lin: Yay!
340 00:54:48.610 ⇒ 00:54:49.779 Miguel de Veyra: Sure everyone can see you.
341 00:54:52.670 ⇒ 00:54:53.800 Amber Lin: Yeah, that’s all from me.
342 00:54:54.650 ⇒ 00:54:55.250 Miguel de Veyra: Yes.
343 00:54:57.110 ⇒ 00:55:17.469 Uttam Kumaran: Cool. So yeah, I think one, we’re gonna do this segment every week, and I will be calling on some other people just on how we’re using AI stuff. Second, if there’s an AI tool that you’ve seen on Twitter or on Youtube, on Google that you want to try, let me know I will get it within reason. Do you want to try it? If you want Chatgpt Pro.
344 00:55:17.680 ⇒ 00:55:22.059 Uttam Kumaran: Use the tool request to request that, and we will get it for you.
345 00:55:22.514 ⇒ 00:55:30.445 Uttam Kumaran: And any, please take a sec to read those, both those notion documents, and make sure that your thoughts and ideas are covered.
346 00:55:31.030 ⇒ 00:55:36.880 Uttam Kumaran: And yeah, this isn’t like a next year thing. This is like a next month thing, by the way. So we should see a lot of this stuff come out
347 00:55:37.170 ⇒ 00:55:39.010 Uttam Kumaran: pretty pretty soon, so
348 00:55:41.190 ⇒ 00:55:44.010 Miguel de Veyra: And I think one other thing I forgot to mention was.
349 00:55:44.180 ⇒ 00:55:59.110 Miguel de Veyra: you can. Actually, I’m not. I’m pretty sure some of you guys have seen it. But you can actually do search and deep search and open AI, which I find very helpful, especially when me and Casey was. We were basically pair programming yesterday on how to work with polytomic
350 00:55:59.490 ⇒ 00:56:00.350 Miguel de Veyra: right?
351 00:56:00.460 ⇒ 00:56:08.559 Miguel de Veyra: And then we found out, you know, hey, it can actually move the files. But deep search search. Yeah, it’s very helpful. I’d advise everyone to use it
352 00:56:09.260 ⇒ 00:56:20.409 Uttam Kumaran: Yeah. Another thing I’m gonna start doing, too, is as I use AI tools. I’m just gonna record a quick loom and send it somewhere. So as you guys are like, if you use an AI tool, you’re like, Oh, my God, this is amazing. Just
353 00:56:20.530 ⇒ 00:56:23.539 Uttam Kumaran: redo it on loom and send it
354 00:56:24.140 ⇒ 00:56:36.480 Uttam Kumaran: like again, I use chat, gpt like on very, very annoyingly. For every single thing and so I wanna make it really clear how everyone can do that. So
355 00:56:37.360 ⇒ 00:56:41.569 Miguel de Veyra: Even messaging in slack. I just Hey, can you rewrite this? Be a bit more professional
356 00:56:41.570 ⇒ 00:56:42.560 Uttam Kumaran: Yeah, I can tell
357 00:56:42.819 ⇒ 00:56:43.080 Miguel de Veyra: Like.
358 00:56:43.080 ⇒ 00:56:49.760 Uttam Kumaran: Tell. We can all tell when it’s written by Chat Gpt. Be more formal
359 00:56:49.760 ⇒ 00:56:51.129 Miguel de Veyra: It will happen
360 00:56:51.590 ⇒ 00:56:56.390 Uttam Kumaran: I’m like, why did you say, Hi, this is Miguel
361 00:56:58.930 ⇒ 00:57:00.800 Miguel de Veyra: It’s a slack message brochure
362 00:57:00.800 ⇒ 00:57:05.436 Uttam Kumaran: It’s like it’s like, dear sir. Thank you so much.
363 00:57:05.900 ⇒ 00:57:08.659 Miguel de Veyra: Them best regards. It’s a it’s a slack message you don’t have to do
364 00:57:08.660 ⇒ 00:57:16.940 Uttam Kumaran: I’m like, dude you didn’t. You didn’t need to run that through. Chat. Gpt right? Now, cool.
365 00:57:17.310 ⇒ 00:57:25.450 Uttam Kumaran: Okay, guys, thank you so so much. If you have any questions, please slack me. And yeah, have a really really great weekend
366 00:57:26.240 ⇒ 00:57:26.920 Miguel de Veyra: Everyone.
367 00:57:27.260 ⇒ 00:57:33.220 Uttam Kumaran: Yeah. Have a good weekend, Hannah and Robert as well. Try not to talk about work too much.
368 00:57:35.570 ⇒ 00:57:37.250 Uttam Kumaran: Okay, bye, everyone
369 00:57:37.250 ⇒ 00:57:38.100 Miguel de Veyra: Thanks everyone, bye.
370 00:57:38.100 ⇒ 00:57:38.490 Nicolas Sucari: I guess
371 00:57:38.490 ⇒ 00:57:39.870 Caio Velasco: Thank you. Bye, bye.