Meeting Title: Data Engineering Community Sync Date: 2025-07-31 Meeting participants: austinW, Uttam Kumaran
WEBVTT
1 00:01:47.330 ⇒ 00:01:48.020 austinW: Yeah, it’s up.
2 00:01:49.020 ⇒ 00:01:49.870 austinW: I think it’s working.
3 00:01:50.100 ⇒ 00:01:54.030 Uttam Kumaran: Yes, sorry for the delay. I’ve literally been in like
4 00:01:54.310 ⇒ 00:02:05.520 Uttam Kumaran: 8 meetings back to back. So this week I you know, I constantly change sort of like how or how I’m doing my stuff. And so Mondays now we have, like all internal.
5 00:02:05.630 ⇒ 00:02:08.940 Uttam Kumaran: and then phase is like, I’m trying to do like all external.
6 00:02:09.070 ⇒ 00:02:15.329 Uttam Kumaran: But then, Wednesday, I’m kind of trying to keep. So I can actually get some like data work done and do some development work. And then.
7 00:02:15.450 ⇒ 00:02:20.780 Uttam Kumaran: then. But then the problem is, when people can’t get to me Wednesday, it just starts to get. They just find a way.
8 00:02:20.780 ⇒ 00:02:25.629 austinW: The separations are good ones, that they’re important. Pardon to catch you on a busy day.
9 00:02:25.630 ⇒ 00:02:31.050 Uttam Kumaran: No, no, you’re totally fine. This is actually what I need to be doing more of. So.
10 00:02:31.050 ⇒ 00:02:31.420 austinW: Again.
11 00:02:31.420 ⇒ 00:02:34.090 Uttam Kumaran: I’m super excited. How do you know, Jody? By the way.
12 00:02:34.270 ⇒ 00:02:40.590 austinW: I ran into Jodi over a discord community that’s stood up by Joe Reese. Joe Reese is an
13 00:02:40.590 ⇒ 00:02:41.459 austinW: oh, yeah. Yeah.
14 00:02:41.460 ⇒ 00:02:45.759 austinW: Data engineering handbook. Yeah. So I’ve been, you know, I guess, kind of
15 00:02:46.080 ⇒ 00:03:03.969 austinW: itching to see more productive professional groups on discord in general. I found it to just be far superior to slack and format, but otherwise that community is really blowing up. And Jody actually tried to put me forward towards one or 2 opportunities he was chasing about 4 months ago. I think it’s his current gig
16 00:03:04.346 ⇒ 00:03:30.670 austinW: but when he put me up to the to the client up against someone with 20 years on paper. They were just like, Let’s go with 20. And then Jody came back to me. Actually, a couple of months later I joined one of his talks, and he was like, Hey, man, I really wish that would have worked out, you know, because I think actually, you would have probably done a little better than this 20 year, Guy, but otherwise he put me into other other opportunities. And at the same time, you know, he suggested, that you and I speak so awesome to talk, to.
17 00:03:30.670 ⇒ 00:03:48.620 Uttam Kumaran: Yeah, what other discord communities do you find are great? Because I mean, I spend a lot of time in the Dbt community when I can. I just joined the measure, which is another like slack community, which is actually really great. But I also find that not a lot of them are great, and I’ve been on the Dbt Slack community for like
18 00:03:48.970 ⇒ 00:03:55.430 Uttam Kumaran: a long time. So the thing about data is, it’s like very circular, like, so.
19 00:03:55.430 ⇒ 00:03:56.070 austinW: Oh, yeah.
20 00:03:56.070 ⇒ 00:04:22.140 Uttam Kumaran: I can’t. I can’t do that. I don’t. I sort of sit on a lot. I was doing a lot of the leadership stuff. And then. Now, I actually am more interested in like new technologies like click house, like some of the new database stuff like Mother Doctv click house code. That’s the stuff I’m like, more about like as more technical. But yeah, interested in. Like, if you have any discord, Rex, like, I would love to join them, and just like read through them.
21 00:04:22.720 ⇒ 00:04:47.010 austinW: No, you should totally get on this one. It’s it’s a practical data. I’ll send you a link out of this one. And again, it being kind of headed by Joe. He’s he’s very agnostic to platforms, that kind of stuff. I can imagine that certain, you know, communities or product based communities would be very insular necessarily. But Joe is very, you know, kind of. He’ll play both sides of it. You know. People want to self promote, but the substance has to be there right?
22 00:04:47.170 ⇒ 00:04:53.531 austinW: Otherwise. I got off the ground with one. I think it’s by the Seattle data Guy Benjamin brokerage in and
23 00:04:54.170 ⇒ 00:05:07.190 austinW: he it was when I was based in downtown Seattle. I was kind of, you know, following the Youtube media of it. And you watch these people. They speak very humbly, you know, about their comeuppance into data and that kind of stuff.
24 00:05:07.360 ⇒ 00:05:14.891 austinW: most of them all kind of going their own anymore. If they’re not, you know, tooting their own horn for doing something crazy at Fang, but otherwise
25 00:05:15.170 ⇒ 00:05:15.900 Uttam Kumaran: Yeah.
26 00:05:16.130 ⇒ 00:05:23.670 austinW: Yeah, no. So no found that this one with Joe is pretty good. They’re reaching about 2,000 members, which I find is pretty exciting. And
27 00:05:24.270 ⇒ 00:05:36.230 austinW: just today they had a good I’ve been doing a book club with them, and they have regular, you know, like events, you know, throughout the week. Necessarily. So there’s some really cool talks. Really cool personalities in that discord.
28 00:05:36.230 ⇒ 00:05:36.820 Uttam Kumaran: Yeah.
29 00:05:37.053 ⇒ 00:05:38.220 austinW: It’s kind of cool. Yeah.
30 00:05:38.350 ⇒ 00:05:49.289 Uttam Kumaran: We have 2 people that I I’ve hired one person off of Dbt. Slack, and I hired another one of our great AI engineers. I found him on the discord for N. 8 N.
31 00:05:49.290 ⇒ 00:05:50.100 austinW: Yeah, yeah.
32 00:05:50.100 ⇒ 00:05:55.879 Uttam Kumaran: I’m like I’m cause for me. I’m like, I’m an engineer. So like I’m I’m where are
33 00:05:56.110 ⇒ 00:06:00.339 Uttam Kumaran: like, where are my people? Right like they’re not on Linkedin. They’re not a lot. A lot of them are linked.
34 00:06:02.210 ⇒ 00:06:13.730 Uttam Kumaran: A lot of the Linkedin is like, maybe the people that like either. Finally bridge the gap, realizing that like, okay, if I want to put myself out there, I have to be on Linkedin. But the real engineers are like they’re.
35 00:06:13.730 ⇒ 00:06:15.410 austinW: Over here biting our fists.
36 00:06:15.410 ⇒ 00:06:16.679 austinW: Yeah, like, where are they like.
37 00:06:16.680 ⇒ 00:06:17.270 Uttam Kumaran: You know.
38 00:06:17.270 ⇒ 00:06:29.779 Uttam Kumaran: So they’re in some discord or slack. So that’s for me. It’s a great filter, and you’re talking. And I tell this to people because I talk to a lot of now I talk to a lot of business people. I’m like the most business person at my company now, which is like pretty sad.
39 00:06:29.920 ⇒ 00:06:41.359 Uttam Kumaran: But I talked to a lot of business people, and and I’m like Dude. I’m I’m hiring off of these places are like what I’m like, dude. Think about it, if you’re on, cause I this what’s where I used to be? If you’re.
40 00:06:41.360 ⇒ 00:06:41.710 austinW: Yeah.
41 00:06:41.710 ⇒ 00:06:47.969 Uttam Kumaran: In this discord or slack of the software you like. You’re a freak, and I want you on my team.
42 00:06:48.288 ⇒ 00:06:57.199 austinW: Freakish or not, you know. People kind of can let their hair down when it’s a persistent, smaller, gated community when it’s kind of everyone’s connections, connections, connections.
43 00:06:57.200 ⇒ 00:06:57.920 Uttam Kumaran: Yes, yes.
44 00:06:57.920 ⇒ 00:07:17.950 austinW: It’s funny, but otherwise it is. It is really great. So no, you should absolutely join this one, and you know you’ll see kind of what goes on. I like it so much, because there are these, you know, very opinionated or very invested. Again, there’s so many writers in that discord. So whether whether they’re going off the deep end of AI things, or if they’re going real book book smart, basically.
45 00:07:17.950 ⇒ 00:07:18.779 Uttam Kumaran: Yeah, yeah, yeah.
46 00:07:18.780 ⇒ 00:07:27.500 austinW: Raising a lot of methodologies, different patterns for warehousing or modeling, whatever it’s just it’s just cool to see. So I just, you know, followed on the phone throughout the day and keeps me freaky. I guess.
47 00:07:27.500 ⇒ 00:07:30.190 Uttam Kumaran: No, in the best way, you know, because.
48 00:07:30.190 ⇒ 00:07:30.610 austinW: Thank you.
49 00:07:30.610 ⇒ 00:07:56.249 Uttam Kumaran: People that are like obsessed with this, because then, you know, for me in my business now, I used to just be an engineer at like product companies. And then, you know, I was leading data and then leading product. But for me, it’s like, I want people that like have a real, you know, interest in like optimizing and like making sure that their skills are getting better right? And so that’s like what great engineers are. And for me. My job is to get tough problems for them.
50 00:07:56.250 ⇒ 00:08:20.369 Uttam Kumaran: And then the more people I can get that have that sort of obsession or interest in growth, our clients benefit that that’s it. And like otherwise, if you get people that clock in and clock out which you know, we’ve worked with those kind of folks. It’s tough in the end. Stakeholder. Whether it is someone internal or our client is the one that suffers because, like they don’t extra distance to be like, is there any different way now? Maybe I did this a couple of years ago, like, let me check, is there any different
51 00:08:20.970 ⇒ 00:08:29.990 Uttam Kumaran: or like, you know, especially now, you can use AI to do research about like what are the best ways I can model. This, you know, some stuff like that is
52 00:08:30.160 ⇒ 00:08:44.249 Uttam Kumaran: is, is like, kind of like, okay, I want to build like the super data team. Because for us, you know, we’re kind of like a decentralized data team. But our clients are each of the, you know, different people that we go after, which is, it’s almost like on hard mode, basically, because.
53 00:08:44.250 ⇒ 00:08:44.820 austinW: Hello! There!
54 00:08:44.820 ⇒ 00:08:51.999 Uttam Kumaran: It’s not like different departments. It’s actually like one layer higher from that right. But I also.
55 00:08:52.000 ⇒ 00:08:52.340 austinW: Right.
56 00:08:52.340 ⇒ 00:09:15.730 Uttam Kumaran: Company is a platform. I don’t. I don’t. I don’t like this different to a lot of consultancies where one client like they don’t have any. They don’t know who the other people are, and other clients, and there’s no camaraderie like that’s not what this is like. We’re not like a bunch of staff like subcontractors. There is like a platform. And but what is a platform? It’s not just like like a technical platform. There is like shared methodologies shared.
57 00:09:15.730 ⇒ 00:09:16.230 austinW: That’s that.
58 00:09:16.230 ⇒ 00:09:40.029 Uttam Kumaran: Shared like architecture shared learning. But also, then we can start building tools for ourselves, like everybody. Okay. But like, that’s a new tool for us. How do we all learn? The best way to Dbt model in cursor? Right? Like that sort of stuff is like, that’s the kind of the platform that I’m trying to build. So anyone that comes in in data can go to the clients benefit. But there is also still like a group. And I think that’s rare.
59 00:09:40.030 ⇒ 00:09:40.460 austinW: And.
60 00:09:40.460 ⇒ 00:09:44.429 Uttam Kumaran: At least, I’ve never seen that in sort of consultancies. You’re typically like.
61 00:09:44.640 ⇒ 00:09:49.690 Uttam Kumaran: whatever your project you’re just like on there. It’s actually way more personal than anything.
62 00:09:49.690 ⇒ 00:10:02.529 austinW: Totally. No, I I see that for me. You know my my own pathway into tech and that kind of stuff. I’ve come from a business degree. And like, it’s really through working into tech teams and understanding the stack and understanding the syntaxes and how databases work.
63 00:10:02.790 ⇒ 00:10:04.040 austinW: You know how relational.
64 00:10:04.040 ⇒ 00:10:04.640 Uttam Kumaran: Get in. How do you.
65 00:10:04.640 ⇒ 00:10:05.000 austinW: Cloud.
66 00:10:05.000 ⇒ 00:10:06.119 Uttam Kumaran: Yeah, like, what was the?
67 00:10:06.120 ⇒ 00:10:10.850 Uttam Kumaran: Because usually, I tell, like, people go come in it from a different like, nobody’s.
68 00:10:10.850 ⇒ 00:10:11.170 austinW: Oh, yeah.
69 00:10:11.170 ⇒ 00:10:26.170 Uttam Kumaran: College, right? So for me, I came in like, I studied computer engineering. But I got a job as a bi analyst engineer. And then, like I both I kind of. I went down, meaning like I was like doing a ton of looker stuff.
70 00:10:26.170 ⇒ 00:10:26.730 austinW: Yeah.
71 00:10:26.730 ⇒ 00:10:28.409 Uttam Kumaran: And I learned airflow and deep.
72 00:10:28.410 ⇒ 00:10:28.760 austinW: It’s a.
73 00:10:28.760 ⇒ 00:10:36.470 Uttam Kumaran: And then I learned python and building like how to host airflow and like sort of went with data engineering. And then I kind of
74 00:10:36.690 ⇒ 00:10:40.950 Uttam Kumaran: went back up, meaning like I kind of went into product. So cause I was so.
75 00:10:40.950 ⇒ 00:10:41.280 austinW: Sure.
76 00:10:41.280 ⇒ 00:10:43.779 Uttam Kumaran: Data products that I was like
77 00:10:43.960 ⇒ 00:10:57.459 Uttam Kumaran: I, when I got the opportunity to build one for my company. I was like, I will use every. I’m really opinionated about Api design integrations like the visual layer of like a analytics product.
78 00:10:57.640 ⇒ 00:11:03.199 Uttam Kumaran: So is there anyone else like, I’ve used so other people’s data products that.
79 00:11:03.580 ⇒ 00:11:04.030 austinW: Right.
80 00:11:04.030 ⇒ 00:11:07.810 Uttam Kumaran: Yeah, just interested, like, how you how you like made that.
81 00:11:08.430 ⇒ 00:11:14.639 austinW: Yeah, no, so totally. You know, I was going into corporate positions since college and in this case.
82 00:11:14.850 ⇒ 00:11:24.539 austinW: you know, it took just a lot of opportunities for growth and and really ambitious growth in a sense. So I got in with SQL. Teams writing, you know, SQL. Doing some procedural stuff.
83 00:11:24.700 ⇒ 00:11:41.450 austinW: After that I got found for a company that was starting to build out like an on-prem data warehouse, and that was they were getting an initial start, and so I had. I got to overlap for about 2 or 3 months with the 2 consultants that were just super book smart. They’re deeply, you know, aware of the SQL stack.
84 00:11:41.550 ⇒ 00:12:07.650 austinW: And then I had to pilot that thing for another 4 years. So that got me book smart on basically data modeling approaches data, warehousing stuff like that. And that’s carried so seamlessly, basically into cloud. And what what data lake is now, or medallion? Or what have you? It was from that company. I was there for a couple of years. I saw them through a pretty big merger, and then I got the opportunity with Amazon, and I stayed there for about 4 years. But not to your point. It was that
85 00:12:10.150 ⇒ 00:12:27.689 austinW: as I was in these companies, you’re drinking that vendor Kool-aid, you’re you’re like, yeah, I was oracle then. Then I was Microsoft, and you put yourself in that box and you realize the proliferation that’s happening around us. And with Amazon. It was like, Oh, well, sweet, this is Fang, but otherwise they burn and churn through the ice, and I was
86 00:12:28.060 ⇒ 00:12:39.090 austinW: by title for way too long and in a sense, but otherwise, I was about the same ways that you know I’m working with my presentation tool. I’m seeing a lot of business value, but I have to go deeper to answer my own questions.
87 00:12:39.090 ⇒ 00:12:45.089 Uttam Kumaran: Want to get to the modeling right? That’s also what I tell my team is that it’s hard to find great analysts, because.
88 00:12:45.090 ⇒ 00:12:45.500 austinW: Tim.
89 00:12:45.500 ⇒ 00:12:49.290 Uttam Kumaran: You’re a great analyst. Then you’re gonna start to poke at the data models and be like.
90 00:12:49.290 ⇒ 00:12:49.770 austinW: Yeah.
91 00:12:49.770 ⇒ 00:12:54.100 Uttam Kumaran: Built these, I might as well build it. And then you also realize it’s more money. So then you
92 00:12:54.100 ⇒ 00:12:59.999 Uttam Kumaran: yeah, you become an analytics engineer right? And I told my team it’s hard to find great analysts, because.
93 00:13:00.570 ⇒ 00:13:06.279 Uttam Kumaran: It’s really a tough job. And I mean for me, I’ve kind of became a full stack data person like I was
94 00:13:06.280 ⇒ 00:13:06.990 Uttam Kumaran: every.
95 00:13:06.990 ⇒ 00:13:16.779 Uttam Kumaran: So yeah, I can still do analysis work, but I also do modeling. I set up warehouses, do Etl, and I like talk to the bit. So for me.
96 00:13:17.210 ⇒ 00:13:22.320 Uttam Kumaran: I like kind of don’t have a particular way. I’m like, what is the problem we’re trying to solve here, you know.
97 00:13:22.540 ⇒ 00:13:36.500 austinW: No, I love. That is because, yeah, you look at that arc, right? You have like data engineering and vie, you have data scientists necessarily, but as soon as you can like really speak at depth to all 3, you know, you’re a different type of team developer at that point, and.
98 00:13:36.500 ⇒ 00:13:36.990 Uttam Kumaran: Yeah.
99 00:13:36.990 ⇒ 00:13:49.759 austinW: You know, there’s such a misnomer, I think, with, you know, new analysts versus old analysts or Bie, you know. Career long, Bie, because, yeah, prior to the move to cloud. Necessarily, we might have just as well been full stack.
100 00:13:49.760 ⇒ 00:13:50.819 Uttam Kumaran: On prem.
101 00:13:50.820 ⇒ 00:14:03.659 austinW: Yeah. So you know, my frustration was, Ssis, my, you know logic was happening within for procedures, necessarily, and we had sequel agents, or what have you but it? You know the tech translates. And that’s that’s the super important part.
102 00:14:03.890 ⇒ 00:14:10.039 austinW: And yeah, I think more and more. You know, I think I also would rather not put myself on one of those titles. But yeah.
103 00:14:10.400 ⇒ 00:14:11.100 austinW: like, no.
104 00:14:11.100 ⇒ 00:14:17.069 Uttam Kumaran: No, but you know the analytics engineer was very recent title. Same data. Science is also a pretty recent title.
105 00:14:17.070 ⇒ 00:14:17.450 austinW: Yeah.
106 00:14:17.450 ⇒ 00:14:21.019 Uttam Kumaran: And like I don’t. I don’t know. I think a lot of what.
107 00:14:21.140 ⇒ 00:14:26.900 Uttam Kumaran: And also again, like sort of Lake House. It’s 1 thing I do like is stuff is getting cheaper, and
108 00:14:27.060 ⇒ 00:14:29.990 Uttam Kumaran: more tools are able to do more of the stack.
109 00:14:30.370 ⇒ 00:14:31.360 Uttam Kumaran: Yeah. So what
110 00:14:31.360 ⇒ 00:14:45.790 Uttam Kumaran: I still think it really matters what you pick to build on, and that’s where like, when I go to clients like, because of how opinionated I was. And I’ve used every vendor. Now I have the joy of like. Now that I have the consultancy I can go get. I can go, get.
111 00:14:45.790 ⇒ 00:14:46.710 austinW: Really don’t open source.
112 00:14:46.710 ⇒ 00:14:48.489 Uttam Kumaran: Product that exists. Right?
113 00:14:48.490 ⇒ 00:14:49.080 Uttam Kumaran: Right? Yeah.
114 00:14:49.080 ⇒ 00:14:56.620 Uttam Kumaran: Great, because you may see on Linkedin, like some product, is like super super great. And then but like, I go try it. I’m like, this is crap. And similarly.
115 00:14:57.830 ⇒ 00:15:02.929 Uttam Kumaran: I find products that nobody has heard of. And I’m like, dude. You guys are criminally underrated.
116 00:15:02.930 ⇒ 00:15:03.760 austinW: Right.
117 00:15:03.760 ⇒ 00:15:06.960 Uttam Kumaran: Or they have way. Better support things like that. And so but then.
118 00:15:06.960 ⇒ 00:15:07.480 austinW: So, yeah.
119 00:15:07.480 ⇒ 00:15:14.340 Uttam Kumaran: Other thing is like Dbt. Is still great, and we use dbt. Core every time, and but I run it for free because I run it on. Get.
120 00:15:14.340 ⇒ 00:15:14.810 austinW: Exactly.
121 00:15:14.810 ⇒ 00:15:21.659 Uttam Kumaran: For free. And yeah, they’re gonna try to move us to fusion. But like, I’m why, that’s not a point of optimization for me. Right? I think.
122 00:15:21.660 ⇒ 00:15:22.330 austinW: Right.
123 00:15:22.330 ⇒ 00:15:26.989 Uttam Kumaran: Writing SQL. In this sort. I think it’s fairly optimized, like I don’t.
124 00:15:26.990 ⇒ 00:15:27.345 austinW: There!
125 00:15:27.700 ⇒ 00:15:28.280 Uttam Kumaran: That’s where.
126 00:15:28.280 ⇒ 00:15:28.620 austinW: Right.
127 00:15:28.620 ⇒ 00:15:29.840 Uttam Kumaran: Fine anymore.
128 00:15:29.840 ⇒ 00:15:32.110 austinW: The compile times are not gonna help you.
129 00:15:32.110 ⇒ 00:15:38.949 Uttam Kumaran: Yeah, like a lot of our clients. We’re we’re running stuff a few times a day like I’m not. They’re dealing with enterprise like.
130 00:15:38.950 ⇒ 00:15:39.470 austinW: Yeah, yeah.
131 00:15:39.470 ⇒ 00:15:43.000 Uttam Kumaran: 100. I’m not in the 100,000 model game, you know I’m in.
132 00:15:43.800 ⇒ 00:15:58.720 Uttam Kumaran: We have hundreds of model game. More of my problem is, how fast can I go from requirement or issue to then creation, pr merge and reflection? And how can I get more over time as we build a stack? There’s.
133 00:15:58.720 ⇒ 00:15:59.050 austinW: Soon.
134 00:15:59.050 ⇒ 00:16:05.140 Uttam Kumaran: Should honestly be less data. Engineering work, more of the workload should shift towards analysis like.
135 00:16:05.140 ⇒ 00:16:06.150 austinW: They’re using them.
136 00:16:06.150 ⇒ 00:16:21.570 Uttam Kumaran: Like when when we come onto a client, it starts off as like we need a data engineer, we need an Ae, right? Let’s say, there’s nothing. So we go on and we implement Snowflake, or we implement something. We then route the day, and we then build the core marts. After that what is the de there for? They’re not so.
137 00:16:21.570 ⇒ 00:16:21.960 austinW: Exactly.
138 00:16:21.960 ⇒ 00:16:49.540 Uttam Kumaran: After our marts are running, too, for maybe 6, 9 months. Okay, like, add columns. But that reduces. So the analyst workload should actually be increasing. Of course, like we have to onboard new domains, new business units, but more of our time should be spent on analysis and helping them make the next smartest decision, right? Because we’re not. We’re not in the like. Okay, cool. Now we have to. We have a hundred that we’re not in that sort of world.
139 00:16:49.540 ⇒ 00:16:54.290 austinW: Right right, not your prize pony in their thing.
140 00:16:54.290 ⇒ 00:16:57.250 Uttam Kumaran: To the business we sell to the top of the business, which means.
141 00:16:57.910 ⇒ 00:17:03.160 Uttam Kumaran: The time we spend optimizing SQL, optimizing data warehouses dude. They don’t care at all. Nobody.
142 00:17:03.160 ⇒ 00:17:03.490 austinW: Are you not.
143 00:17:03.490 ⇒ 00:17:05.649 Uttam Kumaran: In fact, like I don’t. Most of our.
144 00:17:05.650 ⇒ 00:17:06.510 Uttam Kumaran: They don’t want to hear it.
145 00:17:06.510 ⇒ 00:17:19.039 Uttam Kumaran: The people that talk that sign our checks. They don’t. They never asked us who we’re using for data warehouse. Right? I know that if they have a CTO or when they hire a data engineer, that person’s gonna be very happy the way.
146 00:17:19.420 ⇒ 00:17:20.419 austinW: On top. Right?
147 00:17:20.420 ⇒ 00:17:21.000 austinW: Exactly.
148 00:17:21.960 ⇒ 00:17:36.060 austinW: Yeah. And if they don’t have it they don’t have the culture for it. You know. It truly is mentoring the new owner, you know. Hey? Here’s how to keep feed the thing. Here’s how to grow the thing. And yeah, your selection of tools, you know, it’s gonna be extensible and more or less agnostic, you know not.
149 00:17:36.060 ⇒ 00:18:00.740 Uttam Kumaran: And try to get them a good deal and stuff. And so for us, what was really great is like, you know, when I thought about starting this this business and being. And I was like, Okay, I want to be able to make great vendor decisions. But I also know there’s a lot of consultants that are like. We only implement this because they get kickbacks. I said, we’re not taking any kickbacks. All of our materials and stuff. It says, like, we take no kickbacks from any of the vendors. I actually just want to implement the tools that are great because.
150 00:18:00.740 ⇒ 00:18:01.200 austinW: Yeah.
151 00:18:01.200 ⇒ 00:18:05.310 Uttam Kumaran: They make our lives either. The reason they’re great is because they make my life easier like, that’s the function.
152 00:18:05.840 ⇒ 00:18:14.670 Uttam Kumaran: right and secondary. What am I? Gonna they’re gonna give me 10% of like an annual. I don’t. That’s like that’s no amount of money for me to not be agnostic.
153 00:18:14.670 ⇒ 00:18:15.130 austinW: Okay.
154 00:18:15.130 ⇒ 00:18:18.120 Uttam Kumaran: Right? Like, Yeah, that’s not worth the trade off for me. So.
155 00:18:18.120 ⇒ 00:18:18.520 austinW: Right.
156 00:18:18.520 ⇒ 00:18:41.239 Uttam Kumaran: I like. I want the freedom to like, pick and choose what’s in the market and the way I tell my team or my clients, I said, look, I’m in the business of digging holes, and every couple of months I find the best shovel that exists. So I’m constantly looking for the next best Etl tool, and we’re going to try for some clients now, because I’ve heard it’s great from a lot of people, and it’s a it’s it’s way cheaper than a lot of tools.
157 00:18:41.240 ⇒ 00:18:55.530 Uttam Kumaran: Okay, try that. So I’m not like beholden to anyone like whatever. The best solution is is where go for and the 10% of like something is not worth it at all for me to like exclusively implement something, you know.
158 00:18:56.300 ⇒ 00:19:05.800 austinW: No, that’s that’s awesome. And yeah, just it’s super important, because the the proliferation of tools being able to kind of roll your own arrangement of composable tools
159 00:19:05.970 ⇒ 00:19:10.959 austinW: is is critical, and that’s the piece that bypasses everybody that’s been kind of locked into their.
160 00:19:10.960 ⇒ 00:19:11.520 Uttam Kumaran: Yes.
161 00:19:11.520 ⇒ 00:19:30.029 austinW: They’re one company. They’re one consultancy and that’s you know. That’s what kind of put me out is that? You know I was my latest bit was a year with a startup, and it was, you know, truly unplugging the fire hazard that they had all these different platform things, you know, arranged in similar but different, you know permutations of each other as a mess.
162 00:19:30.240 ⇒ 00:19:54.060 austinW: but I had to kind of string out their dependencies. They’re an e-commerce intermediary, and they basically set up a customer to sell their product on Amazon. Ebay, Walmart wish simultaneously. So it was a big pool to kind of jump in and my buddy from Amazon. He was building the team, and I was really the 1st pair of tech hands, and I was listening to your talk with polytonic. And just, you know that experience of being underwater as a tech.
163 00:19:54.060 ⇒ 00:19:54.530 Uttam Kumaran: Yeah, we.
164 00:19:54.530 ⇒ 00:20:08.789 austinW: Necessarily, you know it was that, and you know I took it for a year, and you know I grew some of the primary accounts like 2 X pretty easily. Necessarily so. It was a big growing curve, but otherwise it wasn’t a very healthy pool. You know. They weren’t watering me. If I was a plant.
165 00:20:08.790 ⇒ 00:20:09.409 Uttam Kumaran: Yeah, yeah.
166 00:20:09.663 ⇒ 00:20:18.530 austinW: But otherwise I was seeing these tools kind of go by, and everything, especially with AI. You know, advancements. I just I wanted to be able to put more attention out of stress, basically. So.
167 00:20:18.530 ⇒ 00:20:19.040 Uttam Kumaran: Yeah.
168 00:20:19.040 ⇒ 00:20:22.790 austinW: Yeah, you know, I’ve been kind of just falling with that
169 00:20:23.300 ⇒ 00:20:36.699 austinW: to your point. Trying tools, tools that don’t cost you the arm and the leg don’t have a complicated infrastructure setup that keeps billing you, even even though you thought you got out. And that’s at that point, you know, running at any end, doing a lot of stuff with a llama.
170 00:20:36.700 ⇒ 00:20:37.170 Uttam Kumaran: Cool.
171 00:20:37.432 ⇒ 00:20:53.730 austinW: And just following the release of Llms and just trying to incorporate it, you know again, from the outside, as like a student anymore. So yeah, Jodi’s got me set up with a 6 month contract that should be starting in a week or so. But from that point forward it’s really just, I’m actually back. I’m a Virginia these days, actually.
172 00:20:53.730 ⇒ 00:20:54.690 Uttam Kumaran: Great. Okay.
173 00:20:54.860 ⇒ 00:20:56.359 austinW: That’s where I’m from originally.
174 00:20:57.696 ⇒ 00:21:03.083 austinW: I’m I’m in Centerville right now. But I’m originally from Herndon Reston area. Yeah. So
175 00:21:03.540 ⇒ 00:21:09.719 austinW: yeah, I mean, it’s just a different pool. Seattle kind of dried up. It was really a thing monoculture out there.
176 00:21:09.860 ⇒ 00:21:16.150 austinW: and otherwise, you know, the application pool is just like nailing a paper to the to a board outside, like you might.
177 00:21:16.150 ⇒ 00:21:16.670 Uttam Kumaran: Yeah, yeah.
178 00:21:16.670 ⇒ 00:21:22.330 austinW: Get just piled on, but otherwise, you know, just kind of keeping on with it. And truly.
179 00:21:22.330 ⇒ 00:21:29.610 Uttam Kumaran: Thinking about like career, wise like you’re now kind of on your own like. How do you think? How do you think about the next few years.
180 00:21:30.399 ⇒ 00:21:56.999 austinW: It’s, you know, it’s it’s it’s a lot to rethink about is you know, to. I’d say 2017 to 2019. It was really an unnatural state for companies. You know, money was cheap weren’t here yet, and it was a land grab for developing development talent, right? But then you realize that you know the thing. Managers, you know they can’t, you know, care so much about you because someone’s not caring so much about them. So it’s a very heartless area.
181 00:21:57.000 ⇒ 00:21:57.730 Uttam Kumaran: Totally, totally.
182 00:21:57.730 ⇒ 00:22:05.500 austinW: And so I think that in this case, you know this, this will be my 1st time working on like as a subcontract, effectively for Jodi.
183 00:22:05.710 ⇒ 00:22:17.599 austinW: and to your point is just, I’m so much more interested to being in and out and effective and knowing what I’m seeing, and you know, calibrating the right solve and being that partner for implementation necessarily
184 00:22:18.066 ⇒ 00:22:29.830 austinW: but otherwise, you know, being in this area, it’s to say, what’s the you know, another pool of companies and networking necessarily is Federal contracting something I want to do. I was always kind of
185 00:22:29.980 ⇒ 00:22:43.959 austinW: you know. It wasn’t my taste, but it might be the necessary evil. So right now I’m going for certifications, just so that there’s no question that, you know, even though I’ve been implementing for 13 years or whatever you know. If you need to see this, you know string of letters. It’s there, you know.
186 00:22:44.200 ⇒ 00:22:54.149 austinW: but otherwise, you know, keeping up these communities, keeping up with tools and trying tools. I want to try and no, that you know. What I want out of it is
187 00:22:54.430 ⇒ 00:23:11.400 austinW: it’s it’s not longevity in one place, but it’s it’s always been a capabilities game for me. It’s to say, if I don’t know how to do it. I’m insistent to learn how to do it, and something like networking that’s not going to come overnight, necessarily, or or web is a big deviation necessarily from deep in the data. But
188 00:23:11.510 ⇒ 00:23:20.629 austinW: again, just continue to go T shape and wide, you know. But to be effective is all, because at that point you can prove your value to the right person. Necessarily.
189 00:23:20.630 ⇒ 00:23:21.210 Uttam Kumaran: Yeah.
190 00:23:21.400 ⇒ 00:23:24.939 austinW: As opposing the pledging, pledging allegiance to a Deloitte or a.
191 00:23:24.940 ⇒ 00:23:32.854 Uttam Kumaran: So that’s why I found about this company is like, I just work for a lot of Vc. Backed startups in New York for a while like increasingly smaller ones. And
192 00:23:33.620 ⇒ 00:23:55.029 Uttam Kumaran: this in this game. It’s so much it’s fun. And it’s more stressful. I’d say it’s fun because I don’t play any company. Politics, like our company has no politics where it’s actually, I would say, most people describe it as pretty relaxed company culture. The clients we work for have politics, but because we’re outsiders, we don’t have any requirements. We don’t go to things. We’re.
193 00:23:55.030 ⇒ 00:23:55.430 austinW: All right.
194 00:23:55.430 ⇒ 00:23:58.830 Uttam Kumaran: Their promotion track. We’re not having to have a they’re not a boss.
195 00:23:58.830 ⇒ 00:24:00.196 austinW: Losing days over it. Yeah.
196 00:24:00.470 ⇒ 00:24:09.409 Uttam Kumaran: We don’t do any of that game, you know. And so that’s a real joy. The also, the thing is like in our. So I started this off as a pure data, analytics, consultancy.
197 00:24:09.410 ⇒ 00:24:09.750 austinW: Alright!
198 00:24:09.750 ⇒ 00:24:25.160 Uttam Kumaran: Modern data stack implementations end to end and being partners with companies. But I was using AI in the business for like 2 years. I started this right after, like 3.5 came out. And then, about a year ago, I was like, we’re
199 00:24:25.180 ⇒ 00:24:51.459 Uttam Kumaran: like, I’m sort of running my business like with AI hand to hand. And then I hired an AI engineer, someone who had just found who was like doing Nadn and voice stuff, and I was like, Come, help me just build more stuff for us. So I was working with him directly on the company, and then, a few months later, I was like we should just go offer this as a service like we’re finding out doing. But then, you know, the trouble we had is like for a couple of months. I couldn’t like bridge the gap like
200 00:24:51.460 ⇒ 00:25:10.060 Uttam Kumaran: I was doing Nan, like AI stuff. We were doing kind of common data stuff. But really our next move. And what we’re finding now is like, how do we bridge those right? And so in a couple of different ways, we’re thinking about it. One is like ultimately for data. All of our work ends up as a number on a dashboard. Right? But I always.
201 00:25:10.060 ⇒ 00:25:10.550 austinW: Sure.
202 00:25:10.550 ⇒ 00:25:16.319 Uttam Kumaran: Felt like a dashboard, is like a bad medium for understanding. In fact, the decision.
203 00:25:16.320 ⇒ 00:25:16.950 austinW: But yeah.
204 00:25:16.950 ⇒ 00:25:36.290 Uttam Kumaran: The decision is really like what we’re trying to affect. Right? I talk about how we want clients to make more decisions and more accurate decisions. That’s it. Right? Business. The amount of decisions you can make is is really what the limiting factor is. I don’t know. At that point it gets a little philosophical. I don’t know whether some of our clients really like that. But that’s how I think about even my work.
205 00:25:36.670 ⇒ 00:26:06.099 Uttam Kumaran: So I’m like, okay. But then AI actually should sit as the next layer on top, where it should make the recommendation it should give you like, what is the decision to be made, and then at some point it will start to make those itself. And so that is where I think there’s naturally once we cause, and also the thing about like the reason why I think all these like text to sequel tools. AI data. The problem with them is they have. They don’t have both the semantic understanding and like colloquial business understanding
206 00:26:06.100 ⇒ 00:26:06.890 Uttam Kumaran: absolutely.
207 00:26:06.890 ⇒ 00:26:12.370 Uttam Kumaran: The lovely thing about us is for every client. We have a shitload of Zoom Meetings, slack messages.
208 00:26:12.910 ⇒ 00:26:23.699 Uttam Kumaran: all the Github code, all of our documentation, all the Ddls, all of the Dbt code and the Bi layer. I have way, more documentation.
209 00:26:23.700 ⇒ 00:26:24.900 austinW: Take a pretty good swing on it.
210 00:26:24.900 ⇒ 00:26:35.389 Uttam Kumaran: On what they need that. So I can truly probably build the best data. AI enabled data anything than any off the shelf tool is gonna do right. And so for me.
211 00:26:35.390 ⇒ 00:26:36.080 austinW: Right.
212 00:26:36.080 ⇒ 00:26:50.890 Uttam Kumaran: What I think about is not like, who, what vendor should I come in and plug in, does? It’s like those guys are never gonna get it. They’re gonna be able to solve things like, yeah, like, what are my orders today? But nobody’s asking that. That’s not why they call me, because they can’t question right?
213 00:26:50.890 ⇒ 00:27:13.529 austinW: Yeah, it’s it’s a it’s a really great inflection point. And again, I envy. You know, folks like you that have ventured to do you know the paperwork, and to get the business, and to do the projects, and to be, you know, so many steps in it, you know, because there’s so much more ahead, you know, not to say that, you know. I don’t know that, you know. It’s a path I’ve always, you know, been interested in. But it’s gonna take me working with Jodi’s. And
214 00:27:13.530 ⇒ 00:27:15.010 austinW: yeah, but I also think it’s like.
215 00:27:15.010 ⇒ 00:27:18.550 Uttam Kumaran: It was a it’s. It’s only recently that I can.
216 00:27:18.550 ⇒ 00:27:18.870 austinW: Even.
217 00:27:18.870 ⇒ 00:27:21.690 Uttam Kumaran: Tell you that, like I have this type of dream dude for.
218 00:27:21.690 ⇒ 00:27:22.130 austinW: Sure.
219 00:27:22.130 ⇒ 00:27:28.709 Uttam Kumaran: 2 years, even said even most days, I’m working like a. It’s like, it’s a lot tougher than I thought it was. Gonna be.
220 00:27:29.046 ⇒ 00:27:30.390 austinW: Like. It’s not everything.
221 00:27:30.390 ⇒ 00:27:40.629 Uttam Kumaran: Because people are saying like about Oh, you just turn this on and you get like I think you will, just knowing you through this call you will get like a couple of contracts. You can balance that. But
222 00:27:40.860 ⇒ 00:27:49.579 Uttam Kumaran: turning that into something like a business is just a different thing, and I would think about it before you embark, because it is like a.
223 00:27:49.580 ⇒ 00:27:50.120 austinW: One.
224 00:27:50.120 ⇒ 00:27:55.359 Uttam Kumaran: You can’t go, and you can ask Jody about it because Jody was doing that. And if you call him now, he’s like dude. I’m not doing that.
225 00:27:55.360 ⇒ 00:27:57.290 austinW: It’s a it’s a commitment, you know. It’s.
226 00:27:57.610 ⇒ 00:28:02.490 Uttam Kumaran: Truly, you know, it’s like a second marriage, or whatever else. Again, I’ve been like a I’ve been.
227 00:28:02.800 ⇒ 00:28:07.090 austinW: I’ve been the assistance. Basically, it’s like, I want to keep people on a level head.
228 00:28:07.090 ⇒ 00:28:07.470 Uttam Kumaran: Yes.
229 00:28:07.470 ⇒ 00:28:09.119 austinW: Teams with good culture.
230 00:28:09.568 ⇒ 00:28:17.439 austinW: And yeah, my my interest is not to bear that weight. My interest is to be necessarily, you know, knowledgeable about what I’m doing.
231 00:28:17.440 ⇒ 00:28:18.120 Uttam Kumaran: Bad thing, dude.
232 00:28:18.120 ⇒ 00:28:19.230 Uttam Kumaran: Get the connection here
233 00:28:19.230 ⇒ 00:28:35.470 Uttam Kumaran: like start this because I had some dream of being an entrepreneur for me. To be quite honest, I want to work with smart data people. And I want to choose those people and get those out of my friends. Second, I want it. I want it to go to the next level, which is helping multiple companies simultaneously.
234 00:28:35.470 ⇒ 00:28:35.850 austinW: M.
235 00:28:35.850 ⇒ 00:28:58.630 Uttam Kumaran: Right right, and taking all those learnings and fuel, my curiosity of like, I love learning about business models. We have a flower company as a client. We have an e-commerce, we have cool companies, we have like all these, and so it’s like such a joy to work with all of them and get to build like really great data talent. And so for me, ultimately, the brand is just the broker between the great people and like tough problems.
236 00:28:58.630 ⇒ 00:28:59.350 austinW: Right.
237 00:28:59.350 ⇒ 00:28:59.870 Uttam Kumaran: But the AI.
238 00:28:59.870 ⇒ 00:29:00.370 austinW: It’s really.
239 00:29:00.370 ⇒ 00:29:08.529 Uttam Kumaran: Like the one really unique wrench that I didn’t expect, which was, Oh, maybe we are uniquely advantaged to like
240 00:29:08.730 ⇒ 00:29:22.100 Uttam Kumaran: start doing AI services and then try to marry these 2. But also, you know, you’ll find that a lot of like what’s called like knowledge engineering context, engineers work, dude work. It’s like no different.
241 00:29:22.100 ⇒ 00:29:24.839 austinW: Don’t don’t forget about the basis, and let AI
242 00:29:25.520 ⇒ 00:29:27.530 austinW: for you, as you can understand it.
243 00:29:27.530 ⇒ 00:29:55.199 Uttam Kumaran: It drags all that forward it just to drag the need. Like all the people that started off as like where we just build AI agents. They end up failing because they can’t do the data work. Not only can they not, they can’t get it out, and they can’t measure like for us whenever we take on AI client. I need Kpis. I need to measure what our agents. What are the response? Times like, what we’re running evals like output is a dashboard as well, and that’s because.
244 00:29:55.200 ⇒ 00:29:55.810 austinW: If that’s.
245 00:29:55.810 ⇒ 00:30:03.319 Uttam Kumaran: That’s our baseline like, that’s who we are in. Our DNA is like a data company, you know. So yeah.
246 00:30:03.750 ⇒ 00:30:24.619 austinW: It’s cool. It’s again, it truly is. It’s just unique times, you know. It’s gonna see? But there’s no shortage necessarily of companies in any state of either having nothing, or having some kind of a rat’s nest, or having a great ambition, you know. But again, it takes a good partner consultant, and that’s what that’s what I believe in. More and more is like tiger teams like you just.
247 00:30:24.620 ⇒ 00:30:27.950 Uttam Kumaran: Yeah, yeah, I see, maybe field teams or whatever like.
248 00:30:27.950 ⇒ 00:30:28.320 austinW: Yeah.
249 00:30:28.320 ⇒ 00:30:35.010 Uttam Kumaran: Tell our companies to. I say, when I call them initially, I’m like, throw us into whatever the biggest fire is.
250 00:30:35.010 ⇒ 00:30:36.669 austinW: About your teams. Let’s help them, you know.
251 00:30:36.670 ⇒ 00:30:37.970 austinW: Yeah. And like, I don’t know.
252 00:30:37.970 ⇒ 00:30:38.790 Uttam Kumaran: Because I want to be
253 00:30:38.790 ⇒ 00:30:54.220 Uttam Kumaran: the person, and you’ll find that as we start to solve problems, problems will find us right. We don’t end up with a shortage of work. In fact, we have to do better job of like shortening, like narrowing. And then we have partners that I can pass on the rest of the stuff like you need shopify.
254 00:30:54.220 ⇒ 00:30:54.630 austinW: Tremendous!
255 00:30:54.630 ⇒ 00:31:07.980 Uttam Kumaran: I have a friend that does that. Okay, I have a friend company that does that like so. But I think still, we’re able to make a great recommendation and help in one way or another, versus a lot of consultants will be like, Oh, we don’t do that like that’s out of scope like, I hate that sort of stuff.
256 00:31:08.210 ⇒ 00:31:08.540 austinW: It’s enough.
257 00:31:08.540 ⇒ 00:31:16.710 Uttam Kumaran: Tendency, though of a consultancy, to do that, by the way, like, in order to build something reliable and repeatable, you.
258 00:31:16.710 ⇒ 00:31:20.080 austinW: It’s like when it’s a when is it a boutique? And then when is it a practice area?
259 00:31:20.080 ⇒ 00:31:26.720 Uttam Kumaran: Yeah, exactly. But see, I. So I always push against that. So yeah, the one being like.
260 00:31:27.240 ⇒ 00:31:28.850 Uttam Kumaran: Wait, there’s a way we could do this, though.
261 00:31:28.850 ⇒ 00:31:30.209 austinW: Learnings across all these special.
262 00:31:30.210 ⇒ 00:31:30.750 Uttam Kumaran: Yeah.
263 00:31:30.750 ⇒ 00:31:31.799 austinW: Practices, right now.
264 00:31:31.800 ⇒ 00:31:40.029 Uttam Kumaran: So it’s like, always kind of push against that need to kind of like completely focus on one thing, you know, or focus on one stack.
265 00:31:40.330 ⇒ 00:31:45.119 austinW: If I was hearing from the video correctly. You’ve been at this for a little about a year and a half or so. Is that right?
266 00:31:45.120 ⇒ 00:31:58.719 Uttam Kumaran: Like 2 years. Yeah, like, I incorporated the company. July of 23. I got our 1st client, probably around April of that year, and then that’s the only way I got the money to pay legal to figure everything out.
267 00:31:58.720 ⇒ 00:31:59.340 austinW: Let’s see.
268 00:31:59.340 ⇒ 00:32:02.700 Uttam Kumaran: And we’re about 15 people now, you know. So it’s
269 00:32:02.810 ⇒ 00:32:21.000 Uttam Kumaran: it’s been great. And we’re completely a remote team. We have people across the us, and like everywhere. I wanted to be remote. I’m here in Austin. And so I was like, Let’s just I want to make sure anybody, because I also know that smart people now want to dictate where they are. And so being to go on site and stuff is.
270 00:32:21.090 ⇒ 00:32:35.080 Uttam Kumaran: it’s tough for a lot of people. And I know that like, Hey, I may be able to recruit better people by having that. Not as like a constraint. So a lot of these companies are. Gonna say, come back to office, you know, because they can’t figure out how to communicate on slack and stuff like.
271 00:32:35.340 ⇒ 00:32:58.139 austinW: Well, you know, it’s it’s not an insistence for remote necessarily, but it is. It’s about culture that’s effective for your company when you’re when you’re small and you’re lean and everyone has got, you know, the hygiene to do remote. Well, you know that. And you know everyone’s kind of on a level, you know, with the tech that they’re using. And you can have an amount of mentorship or juniorship around you. And that’s a good, healthy thing. But otherwise, you know, can’t you know it?
272 00:32:58.220 ⇒ 00:33:06.020 austinW: These teams, these consultancy teams it is. You know how how stacked is your your bench, how how good is your roster, and can we knock it out, you know, so.
273 00:33:06.020 ⇒ 00:33:06.370 Uttam Kumaran: Yeah.
274 00:33:06.370 ⇒ 00:33:07.880 austinW: That’s exciting stuff. It’s cool.
275 00:33:07.880 ⇒ 00:33:21.690 Uttam Kumaran: Yeah, man, we should find a way to do some work together. I’m like, I’m really interested in, like how we can do something together, or like I would love for you to even just like meet, you know, one or 2 people on my team and like, keep chatting. Like to be honest, you know, if
276 00:33:22.050 ⇒ 00:33:46.149 Uttam Kumaran: what I try to tell folks is that if your goal is to try to do something like this, or like, figure out what it’s like to run. Something like this company is a great place to do that, because I’m like a complete open book with how it’s going and what the problems are. And I also think we’re going to start to do things in data. And AI that I think we’re on the edge on and I don’t know. I feel like
277 00:33:46.250 ⇒ 00:34:05.749 Uttam Kumaran: my goal is to try to bring on just the best data people to do that. You know, I do think we’re a growing company. We don’t have like a hundred Bajillion dollars, and we don’t have every. But what we do have is really tough problems. And we have a lot of promise to be. I think, one of the better. You know data agencies out there, our goal, like when you ask, like.
278 00:34:05.750 ⇒ 00:34:16.919 Uttam Kumaran: who is our competition. Like, I say, Deloitte, accenture pwcy like I don’t look left to right for inspiration, because I think we’re the only people in our way right now.
279 00:34:17.610 ⇒ 00:34:20.879 austinW: Yeah. You know, bigger team doesn’t make better team necessarily. And.
280 00:34:20.880 ⇒ 00:34:21.250 Uttam Kumaran: Yeah.
281 00:34:21.250 ⇒ 00:34:30.169 austinW: You know it is about that appetite to, you know. Keep, keep up with the technology, make the fit, and then get things kind of Loptite, and we play so many more roles than we do. You know.
282 00:34:30.170 ⇒ 00:34:30.649 Uttam Kumaran: That’s technology.
283 00:34:30.659 ⇒ 00:34:33.859 austinW: All just looking for looking for things. So again, just
284 00:34:34.049 ⇒ 00:34:37.029 austinW: having your you know, your own constraints and capacities.
285 00:34:37.030 ⇒ 00:34:37.400 Uttam Kumaran: Yeah.
286 00:34:37.400 ⇒ 00:34:37.980 austinW: Clear, so.
287 00:34:37.989 ⇒ 00:34:40.809 Uttam Kumaran: Using a lot of AI in the business dude like we are.
288 00:34:40.810 ⇒ 00:34:41.190 austinW: Okay.
289 00:34:41.199 ⇒ 00:34:51.929 Uttam Kumaran: All of our project. Managers use AI to go from meeting transcripts to tickets. We built like an internal platform ui, to manage all of our clients, and, like coded the whole thing.
290 00:34:51.929 ⇒ 00:34:52.349 austinW: I said.
291 00:34:52.350 ⇒ 00:34:57.699 Uttam Kumaran: Built on supa base. We auto transcribe, categorize and score meetings like.
292 00:34:57.700 ⇒ 00:34:58.060 austinW: Awesome.
293 00:34:58.060 ⇒ 00:35:00.610 Uttam Kumaran: We have slack agents like we’re do. We’re just like.
294 00:35:00.610 ⇒ 00:35:01.620 austinW: That’s exciting. Yeah.
295 00:35:01.620 ⇒ 00:35:11.399 Uttam Kumaran: Pushing. I’m pushing AI as much as possible, because what you’ll find is that a lot of our clients, the stuff they ask for is actually way easier than the stuff I ask for.
296 00:35:11.400 ⇒ 00:35:12.849 austinW: Yeah, that’s fair.
297 00:35:12.850 ⇒ 00:35:27.369 Uttam Kumaran: You know, cause I’m like dude. This is possible. We need to build like, for example, we just built an agent that on a schedule, looks at slack, and will tell us messages that haven’t been responded to that like responses that like may have gotten lost.
298 00:35:27.370 ⇒ 00:35:27.730 austinW: Gotcha.
299 00:35:27.730 ⇒ 00:35:34.249 Uttam Kumaran: Problem which is like slack. It does not do well with prioritization or like understanding urgency. So like
300 00:35:34.250 ⇒ 00:35:34.720 Uttam Kumaran: thanks for having
301 00:35:34.720 ⇒ 00:35:49.199 Uttam Kumaran: 2 h, our project manager should get a ping like, Hey, client asked about this, and like, maybe they just missed it right and like. But the next level right, if we have all our slack messages. The AI can probably draft like a half decent message to send.
302 00:35:49.200 ⇒ 00:35:50.659 austinW: And then it starts to.
303 00:35:50.660 ⇒ 00:36:03.829 Uttam Kumaran: Take on like the easy questions, whether it’s even an internal question like, Where is this sop or like? How do I get access to this. Commonly people are blocked just because they don’t get a response. Instead of the AI can start to take that in that
304 00:36:04.220 ⇒ 00:36:12.000 Uttam Kumaran: already used an AI to figure that out. But again, adoption is hard, and so if AI inserts itself, there’s no hiding right? So.
305 00:36:12.000 ⇒ 00:36:12.690 austinW: See it.
306 00:36:12.860 ⇒ 00:36:14.020 Uttam Kumaran: Yeah, that’s the kind it’s like
307 00:36:14.020 ⇒ 00:36:20.769 Uttam Kumaran: like trying to break down and like, just give it a go. You know, I think a lot of consultancies will do that longer term. But
308 00:36:21.030 ⇒ 00:36:23.150 Uttam Kumaran: that’s our edge, you know.
309 00:36:23.470 ⇒ 00:36:51.650 austinW: There you go. I don’t wanna run you along. I’m gonna get you a link to that practical data thing. And we should definitely talk, you know, I’d be happy to keep you up to date. You know how this project’s going. This is. Gonna be like a. It should be a softball pitch type of 6 6 month engagement tools you mentioned click house and you know. Yeah, that tech boggles the mind. Necessarily, you hear about the you hear about the Kafka as being hard to feed.
310 00:36:51.650 ⇒ 00:36:52.070 Uttam Kumaran: Yes.
311 00:36:52.070 ⇒ 00:37:02.929 austinW: Necessarily a company that is doing. Really, I I love their culture. I applied, but I didn’t get in, but they have an awesome product called Tiny Bird. They’re based out of Spain.
312 00:37:03.436 ⇒ 00:37:11.570 austinW: But their CTO is very, you know, just puts gold on the linkedin for what that’s worth, and they push
313 00:37:11.760 ⇒ 00:37:18.106 austinW: really meaningful functionality out of a small team, necessarily. And they’re doing tremendous stuff with.
314 00:37:18.900 ⇒ 00:37:24.370 austinW: you know, I guess spoken language really. Language based Llm requests towards.
315 00:37:24.370 ⇒ 00:37:25.109 Uttam Kumaran: Really, okay.
316 00:37:25.110 ⇒ 00:37:32.040 austinW: Stand up, you know, topics and pipelines. And their their pitch basically is that they take you from
317 00:37:32.410 ⇒ 00:37:51.870 austinW: basically a pub sub topic to an available Api endpoint, you know, pretty damn quick. And they are doing some really cool stuff. They got it. They’ve they’ve got a partnership with Vercel right now. And so what they do is they do a end user analytics. So it’s to say that if you needed to have a live ticker
318 00:37:51.990 ⇒ 00:38:08.080 austinW: going to the front of your website to a million concurrent users. They do. All that streaming and click house is just, you know, phenomenal on that, you know, there’s limitations that you join or manage that stack of data. But otherwise it’s really pretty interesting stuff. So.
319 00:38:08.080 ⇒ 00:38:13.299 Uttam Kumaran: Yeah, and send me send me that discord. By the way, I want to join today. And like, I’ll just probably just binge that tonight.
320 00:38:13.300 ⇒ 00:38:19.170 austinW: There you go. Cool, man, it’s it’s it’s popping off. So yeah, I’ll send you a follow up email, just with that with that link for you. Okay.
321 00:38:19.170 ⇒ 00:38:29.990 Uttam Kumaran: Perfect. Okay, great meeting you. By the way, yeah, keep in touch. I’ll text, let’s let’s chat again sometime for sure. And if I can help with anything or yeah, please let me know.
322 00:38:30.310 ⇒ 00:38:34.380 austinW: Nothing urgent. I’m floating. Got a little bit of work, and we’ll see how it goes. Man.
323 00:38:34.380 ⇒ 00:38:35.580 Uttam Kumaran: Okay, okay, perfect.
324 00:38:35.580 ⇒ 00:38:36.899 austinW: Have a good night.
325 00:38:36.900 ⇒ 00:38:38.360 Uttam Kumaran: Okay. Thank you. Talk soon. Bye.