Meeting Title: Brainforge AI Integration Discussion Date: 2025-08-12 Meeting participants: Uttam Kumaran, Bruno Vendruscolo, Demilade Agboola
WEBVTT
1 00:00:51.690 ⇒ 00:00:52.810 Bruno Vendruscolo: Hello.
2 00:00:55.860 ⇒ 00:00:57.359 Uttam Kumaran: Hey, can you hear me?
3 00:00:58.780 ⇒ 00:01:00.280 Bruno Vendruscolo: I can hear you, how are you?
4 00:01:00.360 ⇒ 00:01:01.779 Uttam Kumaran: Hey, good, how are you?
5 00:01:02.400 ⇒ 00:01:05.050 Bruno Vendruscolo: I’m doing well, I’m doing well. Things are good.
6 00:01:05.250 ⇒ 00:01:07.269 Uttam Kumaran: How’s life since kid?
7 00:01:08.980 ⇒ 00:01:10.660 Bruno Vendruscolo: well, exciting.
8 00:01:12.270 ⇒ 00:01:21.229 Bruno Vendruscolo: busy, crazy, a lot of things happening at the same time, right? So… Olivier’s a full-time job.
9 00:01:23.580 ⇒ 00:01:30.130 Bruno Vendruscolo: And… Yes, a lot of things, a lot of things. She’s growing up so fast, right? …
10 00:01:31.360 ⇒ 00:01:36.459 Bruno Vendruscolo: things are good. Things are good. She’s healthy, we… We love having
11 00:01:37.050 ⇒ 00:01:40.400 Bruno Vendruscolo: Great times with her, just enjoying some time.
12 00:01:41.660 ⇒ 00:01:45.220 Bruno Vendruscolo: And yes, things are overall really good.
13 00:01:46.050 ⇒ 00:01:51.300 Bruno Vendruscolo: Pretty good, really good. It’s a good… it’s a good BZ, except when you cannot sleep.
14 00:01:51.440 ⇒ 00:01:56.270 Bruno Vendruscolo: So, when it happens, which is… Unfortunately.
15 00:01:57.220 ⇒ 00:01:58.290 Bruno Vendruscolo: Frequent.
16 00:01:58.710 ⇒ 00:01:59.320 Uttam Kumaran: Yeah.
17 00:02:00.000 ⇒ 00:02:07.210 Bruno Vendruscolo: This is not good, but besides that, everything else is great. I mean, the family enjoys spending time with her.
18 00:02:07.370 ⇒ 00:02:12.779 Bruno Vendruscolo: Because, … But yeah, all good, all good.
19 00:02:13.010 ⇒ 00:02:24.269 Uttam Kumaran: Nice. Yeah, things on our side are still busy. Like, we’re growing a lot. I think since I last talked to you, we maybe have, like, 5 or 6 more clients that we’re working on.
20 00:02:24.460 ⇒ 00:02:32.800 Uttam Kumaran: We’re doing a lot more stuff around… product analytics, like Amplitude, Mixpanel, we’re working.
21 00:02:32.800 ⇒ 00:02:33.360 Bruno Vendruscolo: Interesting.
22 00:02:33.360 ⇒ 00:02:36.999 Uttam Kumaran: like, we have 3 or 4 new, AI-related
23 00:02:37.360 ⇒ 00:02:41.219 Uttam Kumaran: clients as well, who are implementing AI-related projects.
24 00:02:42.140 ⇒ 00:02:42.900 Uttam Kumaran: …
25 00:02:43.280 ⇒ 00:02:56.690 Uttam Kumaran: But yeah, things are just continuing to grow. It’s been great working with Demolade and a lot of the team, so it’s been… it’s been great. I invited him to this meeting, but I… I think it’s really late his time, so I said if he’s still up, he can join, but …
26 00:02:57.020 ⇒ 00:03:11.419 Uttam Kumaran: Yeah, so it’s been good, dude. Yeah, it’s… we’re still doing a lot of dbt work, but it’s honestly changed. We’re doing, like, some product analytics work, some more analysis work, some dbt work, but then a lot of AI stuff, you know.
27 00:03:12.520 ⇒ 00:03:17.060 Bruno Vendruscolo: Good. Good, and… How has it been? How have you…
28 00:03:18.070 ⇒ 00:03:23.389 Bruno Vendruscolo: how have you been implementing AI for customers? What have you been doing? Like.
29 00:03:24.170 ⇒ 00:03:37.270 Uttam Kumaran: Yeah, so we… so we use a lot of AI in our company, and so we’ve taken a lot of what we learned about where there’s actual ROI in implementing AI, and we go to companies, and it’s a lot of internal optimization.
30 00:03:37.270 ⇒ 00:03:44.280 Uttam Kumaran: So, helping them streamline… like, some of our clients are other consultancies, so streamlining client onboarding.
31 00:03:44.630 ⇒ 00:03:53.940 Uttam Kumaran: Streamlining, you know, just chatting over documents, building, like, RAG systems for people. So it’s kind of case-by-case, but…
32 00:03:54.350 ⇒ 00:04:08.400 Uttam Kumaran: all of our clients just… they don’t have another pathway towards, like, finding, you know, efficiencies in their business, you know, and so… but also, they see AI, but they… they want some help in how they can actually, you know.
33 00:04:08.420 ⇒ 00:04:22.270 Uttam Kumaran: like, go from it being something their board is forcing them to do, or, like, something they’re thinking about is actually something that’s real and driving real change. And so, we come in as, like, consultants, and then ideally, we find one or two pieces to help.
34 00:04:22.360 ⇒ 00:04:32.769 Uttam Kumaran: build on. So we use a lot of N8N, we build, like, agentic workflows for folks, doing a lot of clay for outbound work. It’s kind of a mix of things, yeah.
35 00:04:33.040 ⇒ 00:04:36.329 Bruno Vendruscolo: Okay, this is good, this is good. I had a chance to…
36 00:04:37.520 ⇒ 00:04:39.849 Bruno Vendruscolo: to try things out with N8N.
37 00:04:40.860 ⇒ 00:04:47.369 Bruno Vendruscolo: didn’t go deep into this, because I was mainly looking into some automation.
38 00:04:47.750 ⇒ 00:04:48.740 Bruno Vendruscolo: It’s tough.
39 00:04:49.200 ⇒ 00:04:57.499 Bruno Vendruscolo: that… I could do without an AN, just because it has nodes and an AI,
40 00:04:57.810 ⇒ 00:05:04.030 Bruno Vendruscolo: agent node, it doesn’t mean that I need to use that. So the purpose was not really clear, but I’ve been
41 00:05:04.490 ⇒ 00:05:05.350 Bruno Vendruscolo: …
42 00:05:05.510 ⇒ 00:05:15.849 Bruno Vendruscolo: hearing a little bit more about Clay, and how they can help you find leads, and reach leads, and things like this, so that seems an interesting tool.
43 00:05:16.320 ⇒ 00:05:23.579 Bruno Vendruscolo: I guess a couple of people in Kizuno, they are using clay, for sure. But yes, this is…
44 00:05:25.030 ⇒ 00:05:26.329 Bruno Vendruscolo: This is very interesting.
45 00:05:27.080 ⇒ 00:05:37.459 Uttam Kumaran: Yeah, so we’ve been doing a lot. I mean, in our business, we use it for sales-related stuff, and for all of our clients, you know, we build chatbots for, so we can chat over all their
46 00:05:37.460 ⇒ 00:05:48.640 Uttam Kumaran: documents and meetings and things like that to help our project managers. We’ve built helpful things between, like, meeting happens and, like, getting tickets created and a lot of things there.
47 00:05:49.000 ⇒ 00:05:57.129 Uttam Kumaran: So then we’ve taken a lot of that, and when we go to clients, they have the same problems as us, so it’s pretty easy for us to go and help them, you know.
48 00:05:57.810 ⇒ 00:06:12.119 Bruno Vendruscolo: Yeah, this makes sense. I mean, in my day-to-day, I’ve been using cloud code when developing in dbt, so connecting to shortcut, MCP integrations, right? Notion. So…
49 00:06:12.930 ⇒ 00:06:26.940 Bruno Vendruscolo: I always try to make sure my shortcut cards are very well defined, like objectives, field definitions, and then I just say, hey, hi Claude, can you please help me with this story, please? And it just, like.
50 00:06:27.580 ⇒ 00:06:31.629 Bruno Vendruscolo: It does that magic. So it’s… it’s really nice.
51 00:06:31.630 ⇒ 00:06:41.120 Uttam Kumaran: I literally am… I literally have Cursor in front of me right now. I’m building some dbt stuff today, so… yeah, and everybody on our team uses… is using Cursor for stuff, and…
52 00:06:41.290 ⇒ 00:06:53.129 Uttam Kumaran: It’s sort of, like, the direction we want to head is, like, having everybody we want to develop cursor rules for, for everybody’s, like, repo, and then make sure that everybody can go straight from ticket to, like, something quickly.
53 00:06:53.310 ⇒ 00:06:55.769 Uttam Kumaran: You know, that starts the project off, you know.
54 00:06:56.360 ⇒ 00:06:57.580 Bruno Vendruscolo: Yeah, yeah.
55 00:06:57.910 ⇒ 00:07:04.670 Bruno Vendruscolo: This makes sense. And, I mean, a very well-documented Task, or project?
56 00:07:05.640 ⇒ 00:07:06.500 Bruno Vendruscolo: is…
57 00:07:06.720 ⇒ 00:07:16.430 Bruno Vendruscolo: as important as, like, coding, right? Because tools are doing the code for you, so you really need to ensure, you really need to know what you were doing and why you were doing.
58 00:07:16.430 ⇒ 00:07:19.839 Uttam Kumaran: And when you have everything very, very clear, well-defined.
59 00:07:20.360 ⇒ 00:07:25.670 Bruno Vendruscolo: then things should be easier, because I… SQL is not…
60 00:07:25.900 ⇒ 00:07:38.970 Bruno Vendruscolo: It’s not the actual problem, like, right, the models. It’s always something behind, like a business rule, a logic, that is behind something, the SQL that may not be correct.
61 00:07:39.120 ⇒ 00:07:39.860 Uttam Kumaran: Yes.
62 00:07:39.860 ⇒ 00:07:43.590 Bruno Vendruscolo: So if everything is really clear and defined.
63 00:07:44.380 ⇒ 00:07:46.379 Bruno Vendruscolo: Yeah, SQL is just like SQL.
64 00:07:46.780 ⇒ 00:07:47.550 Uttam Kumaran: Goodbye.
65 00:07:47.780 ⇒ 00:07:48.350 Uttam Kumaran: Yeah.
66 00:07:48.350 ⇒ 00:07:48.950 Bruno Vendruscolo: Yeah.
67 00:07:49.610 ⇒ 00:07:52.599 Uttam Kumaran: So that’s why we’re finding a lot of nice ways between
68 00:07:53.470 ⇒ 00:08:11.949 Uttam Kumaran: the, like, actually, like, planning out projects and using AI to build, but on the data side, it’s all the kind of same old stuff, like, one of the things that we’re trying to do is… is now that we… I’m kind of trying to build ways where the AI can actually help you answer questions, like, it can query your database, it can look at the code.
69 00:08:11.950 ⇒ 00:08:20.639 Uttam Kumaran: And, you know, I think what we’re going to try to do is have AIs almost help with decision-making longer term. And so that’s something that we want to try to…
70 00:08:21.230 ⇒ 00:08:40.969 Uttam Kumaran: tried to explore with clients, which is, like, can we actually deliver, like, a Brainforge agent that maybe it can talk to your database, it can look at your SQL code, and actually start to answer questions about your data, because we have all this semantic knowledge, and we have all the meetings, Notion, like, it can start to get really a lot smarter than, like.
71 00:08:41.059 ⇒ 00:08:45.970 Uttam Kumaran: some of these software that exists that just… they write a SQL query and they give you the answer, you know?
72 00:08:46.560 ⇒ 00:08:55.349 Bruno Vendruscolo: Hmm, well, that’s… That’s a game changer. Yeah, because we were the ones with all the meetings.
73 00:08:55.360 ⇒ 00:09:00.869 Uttam Kumaran: We ask a ton of questions, we have all of our Slack, and all of that we shove into AI.
74 00:09:01.250 ⇒ 00:09:05.759 Uttam Kumaran: And so… If it has all the context, like, it’s getting better, you know?
75 00:09:06.160 ⇒ 00:09:13.820 Bruno Vendruscolo: Yes, all about… all about a project, working with the same client. So if you have all context from meetings, from documentation.
76 00:09:14.210 ⇒ 00:09:20.469 Bruno Vendruscolo: tasks, projects, code, documentation in dbt. If it has access to all of this.
77 00:09:21.010 ⇒ 00:09:24.119 Bruno Vendruscolo: It will very likely know how to answer.
78 00:09:24.510 ⇒ 00:09:26.069 Bruno Vendruscolo: Tricky questions, actually.
79 00:09:26.480 ⇒ 00:09:30.509 Uttam Kumaran: Yeah, yes. But it’s also, like, when did a decision get made?
80 00:09:30.610 ⇒ 00:09:38.070 Uttam Kumaran: Like, or who should go… who should you go ask to about a certain thing? Or maybe even, like, how would you go start to ask a question
81 00:09:38.400 ⇒ 00:09:41.579 Uttam Kumaran: You know, those are all the things that where it’s a huge tax.
82 00:09:41.860 ⇒ 00:09:47.220 Uttam Kumaran: You know, on people. So that’s why, but we want to try to develop some agents for… assist with that.
83 00:09:47.540 ⇒ 00:09:53.070 Uttam Kumaran: And that’s, like, where we’re gonna find some great opportunities between AI and data, I think.
84 00:09:54.400 ⇒ 00:09:57.500 Bruno Vendruscolo: Interesting, interesting. Yeah, looks good, looks good.
85 00:09:57.880 ⇒ 00:10:12.410 Uttam Kumaran: Yeah, dude, so, I mean, we’re starting to grow the team, so I wanted to call you and see what you’re up to, and see if you’re… if you’re… I know last time I sort of talked to you a little bit about, sort of, the company you’re at, but tell me where you’re at, and, like, what you’re curious about, and…
86 00:10:12.800 ⇒ 00:10:25.589 Uttam Kumaran: like, yeah, you’re on the first of my list to kind of call whenever I think about… I’m excited to bring on more data people, so I was like, okay, let me try to call Bruno, let me talk to some other people. So, yeah, let me know where you’re at.
87 00:10:26.170 ⇒ 00:10:32.880 Bruno Vendruscolo: Yeah, this is really nice of you, so, … Again, still working.
88 00:10:33.330 ⇒ 00:10:34.450 Bruno Vendruscolo: A lord.
89 00:10:34.800 ⇒ 00:10:38.029 Bruno Vendruscolo: For Casuno, so starting early.
90 00:10:38.670 ⇒ 00:10:40.889 Bruno Vendruscolo: like, around 6 AM, then I go…
91 00:10:41.450 ⇒ 00:10:48.550 Bruno Vendruscolo: Until, like, 2A, 2 p.m. Then I have a side project that is taking an hour off my day.
92 00:10:49.000 ⇒ 00:10:54.259 Bruno Vendruscolo: … And then I would have a couple of hours.
93 00:10:54.490 ⇒ 00:10:59.020 Bruno Vendruscolo: I think we discussed this before. Yeah. So I would be…
94 00:10:59.230 ⇒ 00:11:07.540 Bruno Vendruscolo: I would still be up for something part-time, like, part-time, like a couple of hours a day that I could spend on a project, if that ever
95 00:11:08.010 ⇒ 00:11:09.649 Bruno Vendruscolo: Could ever make sense.
96 00:11:09.910 ⇒ 00:11:20.270 Bruno Vendruscolo: For this client that I’m working on, not Fusuno, it’s a Brazilian company, so I’m spending every day an hour, on this project. It’s been working because of the
97 00:11:20.800 ⇒ 00:11:29.039 Bruno Vendruscolo: They are not requiring something really fast, they don’t have so much time to spend on this project with me.
98 00:11:29.350 ⇒ 00:11:33.330 Bruno Vendruscolo: And things are going on a… Slow…
99 00:11:33.850 ⇒ 00:11:49.520 Bruno Vendruscolo: motion pace, but it’s… it’s going… it’s going well. We are making progress, we are creating very well curated models, and so on. It’s a very analytics engineering model with dbt and BigQuery, not BI yet.
100 00:11:50.010 ⇒ 00:11:55.320 Bruno Vendruscolo: So it’s working. … And, if there’s something there.
101 00:11:55.800 ⇒ 00:12:02.790 Bruno Vendruscolo: that would require a couple of hours of my day, that could also work. In terms of work, …
102 00:12:04.190 ⇒ 00:12:07.079 Bruno Vendruscolo: I… I’ve been trying to implement AI,
103 00:12:08.580 ⇒ 00:12:16.069 Bruno Vendruscolo: every day, as I told you, so still using VS Code, but with quad code, degrading Notion shortcut.
104 00:12:16.230 ⇒ 00:12:19.750 Bruno Vendruscolo: Learning how these tools Work.
105 00:12:20.330 ⇒ 00:12:24.080 Bruno Vendruscolo: But I haven’t had a chance to stop and think.
106 00:12:24.630 ⇒ 00:12:32.720 Bruno Vendruscolo: How this, … new AI stuff, and all those tools could help actually,
107 00:12:34.140 ⇒ 00:12:47.879 Bruno Vendruscolo: companies create a bigger impact. People are still talking a little bit about automation and things like this. Sometimes they try to implement AI for things that it shouldn’t work. For example, a logistic regression model.
108 00:12:48.350 ⇒ 00:13:04.630 Bruno Vendruscolo: Well, if you want to create a score for a customer, maybe you shouldn’t use AI just to give you a score based on a bunch of metrics. Maybe you should be running a machine learning algorithm, which is something different. I’m also interested in machine learning, so learning more about algorithms, something that is
109 00:13:04.810 ⇒ 00:13:11.390 Bruno Vendruscolo: robust, right? That could help customers, not necessarily using AI, but… Like a real algorithm.
110 00:13:11.710 ⇒ 00:13:12.500 Bruno Vendruscolo: So…
111 00:13:14.210 ⇒ 00:13:21.800 Bruno Vendruscolo: I’ve been also trying to learn a little bit more about data science, machine learning at all, overall.
112 00:13:22.850 ⇒ 00:13:32.419 Bruno Vendruscolo: AI is also interesting, but my day-to-day is basically a lot of analytics engineering, documentation, dbt, data modeling, … Yeah.
113 00:13:32.840 ⇒ 00:13:38.850 Bruno Vendruscolo: Trying to answer quick questions, running some analysis as a sole data person.
114 00:13:39.480 ⇒ 00:13:51.559 Bruno Vendruscolo: But I’m very into everything data, not something very specific. Also had to do a few things with Terraform to manage some services in GCP and all that stuff, so…
115 00:13:51.760 ⇒ 00:13:57.840 Bruno Vendruscolo: It’s a little bit… of everything, but I would say I would love to learn more.
116 00:13:58.040 ⇒ 00:14:07.580 Bruno Vendruscolo: And… And understand how we can create a huge impact on organizational landlord using AI and machine learning.
117 00:14:07.790 ⇒ 00:14:08.530 Uttam Kumaran: Yeah.
118 00:14:08.640 ⇒ 00:14:12.679 Bruno Vendruscolo: That’s what I would be into.
119 00:14:13.640 ⇒ 00:14:19.160 Bruno Vendruscolo: To be honest, I think that… Data transformation, data engineering part.
120 00:14:19.910 ⇒ 00:14:30.169 Bruno Vendruscolo: comes in tandem, right? So those things are happening at the same time. Sometimes they are requirements of all this AI work and machine learning thing.
121 00:14:30.430 ⇒ 00:14:33.090 Bruno Vendruscolo: … But I would…
122 00:14:33.390 ⇒ 00:14:41.529 Bruno Vendruscolo: given that we already have a lot of good data to work on, I would be very interested to learn more about AI and machine learning.
123 00:14:42.110 ⇒ 00:14:42.780 Uttam Kumaran: Okay.
124 00:14:43.110 ⇒ 00:14:49.980 Uttam Kumaran: Okay, that’s helpful to know. I mean, we have some clients that we’re starting where it’s still a lot of dbt work, but I think one thing that
125 00:14:50.020 ⇒ 00:15:00.709 Uttam Kumaran: we’re hoping for is that some people on the data side get interested in some of the AI work, and you can kind of see what that is, and maybe there’s opportunity to collaborate there. I mean, even a lot of the
126 00:15:00.710 ⇒ 00:15:10.169 Uttam Kumaran: the stuff we’re doing. We’re even trying to get everyone to level up on how we’re doing dbt development with AI, and spending more time there.
127 00:15:10.350 ⇒ 00:15:16.200 Uttam Kumaran: how we’re doing, like, dashboard development with AI, and how we, like, use AI in every step of the actual
128 00:15:16.320 ⇒ 00:15:20.719 Uttam Kumaran: you know, analytics engineering process as well, so that could be something interesting, too.
129 00:15:21.380 ⇒ 00:15:30.800 Bruno Vendruscolo: And… Good. This makes sense. And do you think anything could work, actually? Like, given that I…
130 00:15:31.070 ⇒ 00:15:35.880 Bruno Vendruscolo: Could spend a couple of hours a day Wheat base, …
131 00:15:36.590 ⇒ 00:15:40.240 Bruno Vendruscolo: on some client work, do you think this would work?
132 00:15:40.240 ⇒ 00:15:47.450 Uttam Kumaran: Yeah, we have a couple of… so, I mean, there’s kind of a mix, so I’m not… you know, some of our clients are really light on the…
133 00:15:47.630 ⇒ 00:15:56.280 Uttam Kumaran: like, stand-ups and things like that. Some clients, it’s a bit tougher because there’s a project manager, and it’s up to them when the stand-ups are.
134 00:15:56.320 ⇒ 00:16:08.739 Uttam Kumaran: Or if they do grooming and sprint planning. So that’s where it’s tough. Like, if there are those meetings, then I can’t… it’s up to them on, like, what those meetings are. There’s also some clients where it’s, like, really async.
135 00:16:09.030 ⇒ 00:16:15.140 Uttam Kumaran: And it could probably work out. I’ll have to go talk to the team. …
136 00:16:15.280 ⇒ 00:16:22.609 Uttam Kumaran: But also, look, if there’s any opportunity even to consider you for larger engagement or full-time, you let me know.
137 00:16:22.940 ⇒ 00:16:31.369 Uttam Kumaran: Like, that’s where I think we are keep… we’re continuing to grow both our analytics engineering team, but again, like.
138 00:16:31.750 ⇒ 00:16:48.379 Uttam Kumaran: we’re so forward on the AI piece, so that whatever you’re doing on AI is here, you know, we’re doing so much that I think you have a good home to do a lot of that, and to start to also see what it’s like to deploy NA to N at scale.
139 00:16:48.380 ⇒ 00:17:00.519 Uttam Kumaran: to actually deploy a lot of AI workloads. So that could be really interesting as well, where I think even if that’s something, you know, you want to get into long-term, there’s a lot of work we’re doing there. And then ultimately, like.
140 00:17:00.520 ⇒ 00:17:17.029 Uttam Kumaran: for me, I’m looking for partners on our data team on how we bridge the two together. Like, how do we start using AI in every step of the development process, from ticket to scoping, to then using Cloud Code or Cursor to write the first version, AI for PR reviews.
141 00:17:17.050 ⇒ 00:17:33.069 Uttam Kumaran: And then finally, using AI to actually help us even, like, triage metaplane logs or dbt errors, like, as many places as we can use it is what we’re thinking. So I think that’s a fun project, you know, for folks that have sort of been in this world for a bit.
142 00:17:33.310 ⇒ 00:17:50.639 Uttam Kumaran: You know, so that’s… for me, like, because I’m running the company, I’m, like, putting… as far as I can push it, wherever we can push it, it’s… there’s… you have… you have sort of my blessing, so we’re trying everything we could do to try to run this business more efficiently, but also, like, have our team work on…
143 00:17:50.830 ⇒ 00:17:54.900 Uttam Kumaran: Actually, more interesting things, and, like, spend time with the clients more, you know?
144 00:17:55.930 ⇒ 00:18:00.850 Bruno Vendruscolo: This makes a lot of sense, yes, I’m… this… This sounds great. …
145 00:18:02.040 ⇒ 00:18:03.820 Bruno Vendruscolo: This is really good, this is really good.
146 00:18:04.000 ⇒ 00:18:06.220 Bruno Vendruscolo: … Well…
147 00:18:07.690 ⇒ 00:18:14.969 Bruno Vendruscolo: because Runo is taking basically all my time, right? So, it’s… it’s my full-time job, besides Olivia.
148 00:18:14.970 ⇒ 00:18:16.810 Uttam Kumaran: Yeah, yeah, yeah. It’s my full-time job.
149 00:18:16.820 ⇒ 00:18:18.810 Bruno Vendruscolo: But, ….
150 00:18:19.080 ⇒ 00:18:34.030 Uttam Kumaran: But maybe if there’s even, you know, even some of these projects that, like, look, if it’s not a client, but maybe if we need some help, like, for example, it would be great to work with someone specifically on just, like, how we can start to use AI in the dbt development process.
151 00:18:34.150 ⇒ 00:18:35.539 Uttam Kumaran: Right? So that’s, like…
152 00:18:36.030 ⇒ 00:18:51.060 Uttam Kumaran: generating, like, SOPs on, like, how to use… how to use Cloud Code or Cursor to write dbt models, choosing the best vendor for PR reviews for using AI. That’s something that, look, maybe it’s not client work, but it’s… we’re… we would be the client.
153 00:18:51.060 ⇒ 00:19:01.730 Uttam Kumaran: Right? Where you would work directly with me and maybe one other person on, like, hey, how do we transform our development process using AI? It’s kind of an interesting meta-project, you know, about, like.
154 00:19:01.760 ⇒ 00:19:08.909 Uttam Kumaran: hey, we have several clients we’re doing analysis work for, that we’re doing DBT work for, that we’re doing…
155 00:19:08.970 ⇒ 00:19:13.499 Uttam Kumaran: data engineering work for, and how can we start to insert AI into each of those processes?
156 00:19:13.710 ⇒ 00:19:15.720 Uttam Kumaran: Could be an interesting project that’s, like.
157 00:19:15.910 ⇒ 00:19:31.469 Uttam Kumaran: it’s not… it’s not necessarily, like, cookie cutter, but it is, like, kind of an interesting thing to work probably directly with… with Neon. You know, where… something that we could have definitely accomplished with, like, a couple hours of work, you know? And it gets… it gets you out of, like, the…
158 00:19:31.640 ⇒ 00:19:33.300 Uttam Kumaran: Day-to-day client stuff.
159 00:19:33.620 ⇒ 00:19:39.279 Uttam Kumaran: But it also is, like, still kind of like… it’s almost like kind of a research-type thing, you know, where…
160 00:19:39.440 ⇒ 00:19:43.760 Uttam Kumaran: I’m really curious on how you’re using AI to write, to do a lot of AE work.
161 00:19:43.940 ⇒ 00:19:49.879 Uttam Kumaran: And then for us, I want to scale that into, like, everybody who does data at Brainforge, I want to be AI-enabled.
162 00:19:51.220 ⇒ 00:19:59.629 Bruno Vendruscolo: Well, that sounds great. If this is something that you really want to implement, like this internal type of work, an internal project.
163 00:20:00.010 ⇒ 00:20:02.680 Bruno Vendruscolo: … I would love to be part of.
164 00:20:02.910 ⇒ 00:20:03.590 Uttam Kumaran: Okay.
165 00:20:04.070 ⇒ 00:20:07.479 Bruno Vendruscolo: could help however I can, could also learn from you.
166 00:20:07.670 ⇒ 00:20:16.309 Bruno Vendruscolo: That’s… that’s amazing, and … like client stuff. I love client stuff, this is interesting.
167 00:20:16.570 ⇒ 00:20:23.940 Bruno Vendruscolo: But I mean, working for an internal team is also something really amazing in the end.
168 00:20:24.320 ⇒ 00:20:26.880 Bruno Vendruscolo: … So….
169 00:20:26.880 ⇒ 00:20:33.160 Uttam Kumaran: I don’t know if you work with us for a bit, then if you want to join it full-time, you can decide on your own time.
170 00:20:33.360 ⇒ 00:20:34.040 Bruno Vendruscolo: Yeah, and I mean….
171 00:20:34.040 ⇒ 00:20:46.870 Uttam Kumaran: So I, like, for me, you know, I want to take advantage of folks that are in the weeds doing data work, but kind of see the impact that AI is having. And then for me, I want to scale that to all of our engineers.
172 00:20:47.080 ⇒ 00:20:58.330 Uttam Kumaran: Right? And start to work on more complicated things, like, for example, one idea I had is, like, can we have AI take a screenshot of a dashboard every day, and sort of produce
173 00:20:58.560 ⇒ 00:20:59.830 Uttam Kumaran: an analysis.
174 00:21:00.180 ⇒ 00:21:00.910 Uttam Kumaran: Right?
175 00:21:00.910 ⇒ 00:21:01.470 Bruno Vendruscolo: Yeah.
176 00:21:01.470 ⇒ 00:21:04.079 Uttam Kumaran: That’s something that, like, that’s like an analyst’s job.
177 00:21:04.270 ⇒ 00:21:07.350 Uttam Kumaran: Right? But can we start to build something that helps us do that?
178 00:21:07.860 ⇒ 00:21:13.060 Uttam Kumaran: Yeah, so we have some interesting projects like that where we would be doing things that I don’t think any other
179 00:21:13.130 ⇒ 00:21:31.439 Uttam Kumaran: consultancy is thinking about, and it would allow us to go above and beyond for our clients within their budget, and so that’s sort of the stuff that’s like, okay… But also, it’s like, okay, cool, how should we structure our tickets so it’s easily able to move into the cursor?
180 00:21:31.520 ⇒ 00:21:35.459 Uttam Kumaran: For work, right? Like, even just having a guideline on that.
181 00:21:35.490 ⇒ 00:21:53.299 Uttam Kumaran: And then allow… showing our PRs, like, showing our PMs, hey, all your tickets that are data modeling, they need to be structured in this way. And then you also go to the AEs, and you’re like, hey, you need to have your linear MCP set up, you need to be able to bring in your ticket like this, here’s how to use Claude, here’s how to use Claude, or whatever, to do your planning step.
182 00:21:53.920 ⇒ 00:21:57.519 Uttam Kumaran: review that, and then have it execute, right? That’s… those are the stuff where
183 00:21:57.640 ⇒ 00:22:03.349 Uttam Kumaran: like, I just don’t have the time to sort of go and sit with everybody and, like, think about what is our…
184 00:22:03.680 ⇒ 00:22:07.090 Uttam Kumaran: Sort of platform documentation for how to develop
185 00:22:07.850 ⇒ 00:22:10.219 Uttam Kumaran: how to develop with AI, right?
186 00:22:10.450 ⇒ 00:22:10.980 Bruno Vendruscolo: Yeah.
187 00:22:11.110 ⇒ 00:22:12.839 Bruno Vendruscolo: Yes, this makes a lot of sense.
188 00:22:13.260 ⇒ 00:22:17.139 Bruno Vendruscolo: Sounds like a lot of work, but very interesting work.
189 00:22:17.140 ⇒ 00:22:17.680 Uttam Kumaran: There you go.
190 00:22:18.300 ⇒ 00:22:20.700 Bruno Vendruscolo: I mean, you could learn a lot from this, right?
191 00:22:20.700 ⇒ 00:22:22.589 Uttam Kumaran: You’re done. Hey there, Milate.
192 00:22:22.970 ⇒ 00:22:25.310 Demilade Agboola: Hey, I was going to say that, like.
193 00:22:25.490 ⇒ 00:22:28.489 Demilade Agboola: I know Bruno well enough to know that he’s quite excited.
194 00:22:28.490 ⇒ 00:22:29.105 Bruno Vendruscolo: Oh….
195 00:22:31.040 ⇒ 00:22:34.789 Uttam Kumaran: I don’t know how long you’ve been here, I don’t know how much you heard, but yeah, I was kind of like, okay.
196 00:22:35.200 ⇒ 00:22:44.089 Uttam Kumaran: Well, if you’re gonna only be available a few hours, like, well, we kind of have this AI thing that I feel like we haven’t really been able to press on, so…
197 00:22:44.430 ⇒ 00:22:46.490 Uttam Kumaran: Could be something interesting.
198 00:22:47.520 ⇒ 00:22:56.090 Demilade Agboola: Yeah, I mean, also, I think one of, and Bruno did mention it earlier, one of Bruno’s super strengths is, like, he’s really good at documentation,
199 00:22:56.400 ⇒ 00:23:00.010 Demilade Agboola: And, like, I know that in terms of…
200 00:23:01.150 ⇒ 00:23:10.260 Demilade Agboola: working on stuff and being able to ensure that people come after and read and understand what the thought process is, what the end goal is, and all of that. Like, Bruno’s really good at that.
201 00:23:10.530 ⇒ 00:23:14.500 Demilade Agboola: So yeah, I do know that…
202 00:23:14.650 ⇒ 00:23:26.990 Demilade Agboola: I’m not exactly sure what capacity it will be able to, like, find the perfect fit, but I do know that, definitely anything that Bruno comes into, like, working on, especially, like, the internal stuff.
203 00:23:27.310 ⇒ 00:23:33.199 Demilade Agboola: In terms of just, like, creating playbooks and, SOPs.
204 00:23:33.930 ⇒ 00:23:45.240 Demilade Agboola: For different things that we can do within the company. I do know, like, he’ll come through and… but we’ll probably get a lot of them up, you know, pretty, pretty quickly.
205 00:23:46.690 ⇒ 00:23:49.739 Bruno Vendruscolo: No, I mean, I’ll start calling you, Demilade, for my interviews.
206 00:23:49.740 ⇒ 00:23:55.990 Demilade Agboola: We all need a wingman in interviews. Maybe that should be a thing, you know?
207 00:23:57.100 ⇒ 00:23:58.849 Demilade Agboola: Neither of you.
208 00:24:01.620 ⇒ 00:24:03.550 Demilade Agboola: But, yeah, ….
209 00:24:04.190 ⇒ 00:24:15.229 Uttam Kumaran: We talked about this, but it’s like the time, because I would say one of the big challenges of this business is time working in the business and time working on the business, right?
210 00:24:15.230 ⇒ 00:24:15.620 Demilade Agboola: Yeah.
211 00:24:15.620 ⇒ 00:24:30.989 Uttam Kumaran: challenge, and I would say a lot of our time, especially demo last time, is going to working on the clients that he’s on, and I… we… but I know, demo, you can say it’s always been a clear thing that we’ve tried time and time again to build some of this platform, you know? Like, how do we…
212 00:24:31.010 ⇒ 00:24:37.620 Uttam Kumaran: how do we commonly execute on these? What’s our common perspective on the way we do modeling work, pipeline work?
213 00:24:37.650 ⇒ 00:24:54.490 Uttam Kumaran: But it’s not something that I’ve ever… we… because all the smart people, we typically have working on clients, so you never get a chance to sort of, like, step away from that. And I think it’s something that we will… over time, I think everybody will have some free time to work on that, but…
214 00:24:54.560 ⇒ 00:25:02.129 Uttam Kumaran: Right now, it’s so busy that, you know, if maybe that is… it’s sort of a blessing in that, like, hey, if you can only do a couple hours.
215 00:25:02.270 ⇒ 00:25:16.270 Uttam Kumaran: whenever, and then maybe it is, for example, building out our, like, SOPs for how we tackle the common data problems. And then, within those SOPs, it’s like, how does AI fit in to execute this, right? So, like, for example, what are the most common data tickets?
216 00:25:16.270 ⇒ 00:25:23.879 Uttam Kumaran: investigate a metric that’s off, right? Okay, so here’s, like, here’s a structure of how we would do that.
217 00:25:23.940 ⇒ 00:25:37.070 Uttam Kumaran: in a world without AI. Now, if you have AI, here’s the best way for, if a client comes to you and says, a number is off, you should use cursor or whatever to go diagnose, right? And here’s the new path
218 00:25:37.090 ⇒ 00:25:43.249 Uttam Kumaran: to doing that. And starting to build those ways where it’s, like, almost, like, how do we start to do the common tasks
219 00:25:43.260 ⇒ 00:25:58.939 Uttam Kumaran: with AI as a core part of it. And similar, like, another thing is, like, it’s on our backlog right now, is it’s on our team to choose, a tool for PR reviews, right? What is the best AI PR review? How can we customize it so
220 00:25:59.050 ⇒ 00:26:17.420 Uttam Kumaran: when we ship a new PR that’s for a new data model, the AI actually does a great review of it. We’re kind of looking at a couple different vendors, trying to pick the best one, have to customize it a little bit. There’s things like that, you know? Another thing is, like, hey, you have a data modeling task, you have to create a new dimension table.
221 00:26:17.500 ⇒ 00:26:21.640 Uttam Kumaran: Okay, but, like, how… in a world of AI, like, how should you start that?
222 00:26:21.730 ⇒ 00:26:25.990 Uttam Kumaran: Right? Versus just, like, opening it up and sort of starting from a blank page.
223 00:26:26.160 ⇒ 00:26:33.889 Uttam Kumaran: What’s the best way? Maybe you take the meeting notes, you then take the ticket, you go into Kirscher, you have a plan.
224 00:26:34.110 ⇒ 00:26:38.789 Uttam Kumaran: edit the plan, right? Like, it’s like, what are these new AI SOPs for a data team?
225 00:26:40.980 ⇒ 00:26:42.409 Bruno Vendruscolo: Always makes a lot of sense.
226 00:26:43.760 ⇒ 00:26:49.330 Uttam Kumaran: And then the nice thing is, you have immediate feedback, because everybody will try it and can tell you.
227 00:26:49.330 ⇒ 00:26:49.850 Bruno Vendruscolo: Yeah.
228 00:26:49.850 ⇒ 00:26:50.640 Uttam Kumaran: Yeah.
229 00:26:52.280 ⇒ 00:27:02.170 Bruno Vendruscolo: Yes, I mean, honestly, look, … I’m always around, so… For now, if you…
230 00:27:05.200 ⇒ 00:27:13.000 Bruno Vendruscolo: If you just want me to be part of this, like, community, or somehow start helping with anything at all, …
231 00:27:14.390 ⇒ 00:27:18.139 Bruno Vendruscolo: … looking into some stuff, I can…
232 00:27:19.070 ⇒ 00:27:27.920 Bruno Vendruscolo: We could try to start something very flexible and informal, if you want to. And then we see how this evolves.
233 00:27:29.050 ⇒ 00:27:41.260 Uttam Kumaran: That’s how we do things, like, I just want to… I think it’s, for me, it’s like, we have an interesting set of problems. I think you have someone like me that’s trying to invest there, but what I lack is the time.
234 00:27:41.500 ⇒ 00:27:54.699 Uttam Kumaran: But what I don’t lack is the vision, and I know what’s possible in AI, you know? So, I think we also have great people that are working on clients that I know could benefit a lot from this, and it would directly hit the
235 00:27:54.940 ⇒ 00:27:59.160 Uttam Kumaran: You know, the time we spend, the ability to work on more clients, like, it’s a great investment.
236 00:27:59.400 ⇒ 00:28:06.699 Uttam Kumaran: for the company, but it needs to be one more step beyond, like, everybody use cursor, right? There has to be a couple of different things.
237 00:28:08.600 ⇒ 00:28:23.770 Uttam Kumaran: And also, like, even for our project managers, right, commonly, like, they want some help on how to structure these projects, what questions to ask clients, like, what are the common roadmaps, like, just having another person that is adept at DBT and AE work.
238 00:28:24.570 ⇒ 00:28:30.970 Uttam Kumaran: … And is not swamped in client stuff, could be… could be helpful. …
239 00:28:31.150 ⇒ 00:28:35.510 Uttam Kumaran: You know, so it… that’s kind of… could be a good place to start.
240 00:28:36.410 ⇒ 00:28:39.239 Bruno Vendruscolo: Yes, yes, I agree. And…
241 00:28:40.730 ⇒ 00:28:43.869 Bruno Vendruscolo: However you prefer to… to manage and….
242 00:28:44.020 ⇒ 00:28:46.080 Uttam Kumaran: Sure. And start this, ….
243 00:28:46.080 ⇒ 00:28:47.570 Bruno Vendruscolo: Have access to my mail.
244 00:28:47.980 ⇒ 00:28:54.630 Bruno Vendruscolo: if you want to give me access to his live channel, if we just want to keep chatting, … I mean, I’m…
245 00:28:55.350 ⇒ 00:29:00.790 Bruno Vendruscolo: 100% open. Sure. And totally available. So…
246 00:29:01.670 ⇒ 00:29:03.669 Bruno Vendruscolo: If this makes sense to you.
247 00:29:04.020 ⇒ 00:29:11.390 Bruno Vendruscolo: the overall idea makes sense to me. Sounds like a project that could bring a lot of value for the team and clients.
248 00:29:11.570 ⇒ 00:29:14.030 Bruno Vendruscolo: Right? So… yeah.
249 00:29:15.230 ⇒ 00:29:29.420 Uttam Kumaran: Yeah, so maybe I’ll… I’ll put some more thought into it, and I’ll send you an email, and I’ll see, because ultimately, I want to pair you maybe with one other… one or two other people that can sort of give you all the perspective needed, and then we can come up with some milestones, like…
250 00:29:29.550 ⇒ 00:29:37.520 Uttam Kumaran: Whether it’s some documentation, or whether it’s some sort of measurement, like, hey, everybody who’s working on data, clients are all using
251 00:29:37.890 ⇒ 00:29:42.270 Uttam Kumaran: the AI systems, we can measure that. I don’t know, we’ll think… we can come up with something.
252 00:29:44.080 ⇒ 00:29:52.010 Uttam Kumaran: But, yeah, maybe… I think that could be something good, you know. Now, especially hearing that you have… you’re now using Cloud Code, and you kind of get it.
253 00:29:52.290 ⇒ 00:29:59.359 Uttam Kumaran: It’s like, the next level is, okay, how… if you were to be running a data team, how do you get everybody on the data team
254 00:29:59.510 ⇒ 00:30:05.850 Uttam Kumaran: To use this, and what are the… and, you know, one of the challenges, like, what are the common situations in which to use which
255 00:30:05.980 ⇒ 00:30:17.420 Uttam Kumaran: what… what method, you know? So… so yeah, maybe let me… let me think about it this week and get back to you, and then… yeah, totally. And then we also… we also have, like, a Friends of Brain Forge chat in Slack.
256 00:30:17.600 ⇒ 00:30:29.760 Uttam Kumaran: Where we just have, like, a lot of guests, actually, where we just… there’s people that are talking, so I can just already add you there. We send AI stuff, and I sort of send ideas there, too, but yeah, I think we could probably do something, for sure.
257 00:30:30.640 ⇒ 00:30:33.769 Bruno Vendruscolo: Yeah, this could work. So Slack could work, email could work.
258 00:30:33.890 ⇒ 00:30:38.629 Bruno Vendruscolo: We could start a draft, maybe even a Notion page, to start brainstorming.
259 00:30:38.630 ⇒ 00:30:48.069 Uttam Kumaran: I’m gonna bring them some stuff, and so another thing that we’re working on, and I’ll share this with you, is we’re working on sort of our AI platform
260 00:30:48.250 ⇒ 00:30:57.249 Uttam Kumaran: architecture. So, this is something that, Sam on our team started working on. But basically, like.
261 00:30:57.860 ⇒ 00:31:15.299 Uttam Kumaran: this is a document about just, like, all of the different processes that we’re using AI in the company right now. And so, for example, like, for me, I gave the sort of broader business vision, which is, like, what are the goals, right? So, for me, it’s, like, leverage per employee, client NPS, and…
262 00:31:15.350 ⇒ 00:31:29.349 Uttam Kumaran: out… out-compete competitors. But he’s… here are also, like, the core KPIs, right? So, like, I want to look at, like, how much time are people sending in meetings, how many Slack messages, revenue per employee, but also, like, tokens used, workflows called.
263 00:31:29.400 ⇒ 00:31:40.960 Uttam Kumaran: Like, things like that. You know, so we’re starting to… like, these are all the processes by which we’re starting to map out in the company, and how AI is going to be used
264 00:31:41.120 ⇒ 00:31:42.669 Uttam Kumaran: to accomplish them.
265 00:31:43.380 ⇒ 00:31:44.060 Uttam Kumaran: …
266 00:31:44.250 ⇒ 00:31:51.800 Uttam Kumaran: And so there’s a clear piece here with… for engineering, which is, like, we only have a couple of here, but we could start to build this up, which is, like.
267 00:31:52.400 ⇒ 00:31:58.620 Uttam Kumaran: how do… how does our data team… what are all the processes in our data team, and how are they being augmented with AI?
268 00:31:58.800 ⇒ 00:32:02.069 Uttam Kumaran: And we have all the tools here that we’re using.
269 00:32:02.330 ⇒ 00:32:08.609 Uttam Kumaran: Like, all their repos for stuff, but that’s gonna be a great place, is to, like, just take this section and
270 00:32:08.820 ⇒ 00:32:16.409 Uttam Kumaran: And we can kind of work together on, like, okay, how do we build a data… an AI-native data team, right? If you were to re… think about everything.
271 00:32:16.970 ⇒ 00:32:30.229 Uttam Kumaran: all options are on the table, how do we get everybody on the existing team to be AI-enabled? But then also, how does the next Brainforge person, when they come in and they join, like, Bruno, if you were to join, you knew nothing of AI,
272 00:32:30.290 ⇒ 00:32:47.730 Uttam Kumaran: how do we… how do we get you onboarded, get you to see the benefit, and also start to use it? Because one of the challenges is we’re not going to find data people that are using AI stuff. Everybody’s just doing their job, like, normally. It’s… it’s… it’s a lot to learn, and also a lot of companies aren’t, like, promoting it, so…
273 00:32:47.790 ⇒ 00:32:54.040 Uttam Kumaran: unless you’re, like, figuring it out on your own, I want this to be like, hey, when you come join us, you get supercharged.
274 00:32:54.100 ⇒ 00:33:08.570 Uttam Kumaran: Right? You… it’s like, you get, like, you literally get, like, a HGH, you get steroids, like, when you come join us as a data person, because we have these great cursors, we have great cursor methods, we have these great SOPs on how to use AI to plan work.
275 00:33:09.070 ⇒ 00:33:09.800 Uttam Kumaran: …
276 00:33:09.970 ⇒ 00:33:17.090 Uttam Kumaran: But the nice thing is, like, as I showed here, we have PM, operations, finance, the only angle you would have is engineering.
277 00:33:17.240 ⇒ 00:33:25.469 Uttam Kumaran: Because I’m already thinking about other stuff, Amber on our team is thinking about project management. So you… but you can see, systematically, I’m going through the company.
278 00:33:25.710 ⇒ 00:33:28.739 Uttam Kumaran: and we’re finding the ways that we’re using AI everywhere.
279 00:33:29.230 ⇒ 00:33:29.930 Uttam Kumaran: …
280 00:33:30.630 ⇒ 00:33:43.999 Uttam Kumaran: So that’s… that would sort of be… and also, I don’t know if I mentioned, we also have an AI team, so it’s not like there’s not people thinking about this. They’re also working on clients, though, so they’re also very busy. But, like.
281 00:33:44.450 ⇒ 00:33:50.150 Uttam Kumaran: how to do something, if we want to build our own MTP servers, everything is, like, totally on the table.
282 00:33:50.310 ⇒ 00:33:54.479 Uttam Kumaran: You know, so I think it’s an interesting project.
283 00:33:56.030 ⇒ 00:33:58.630 Bruno Vendruscolo: I mean, it is, it is for sure. …
284 00:33:59.440 ⇒ 00:34:04.090 Bruno Vendruscolo: And this is also what I’ve been trying… actually, I’ve been realizing.
285 00:34:04.410 ⇒ 00:34:08.320 Bruno Vendruscolo: how this… how this actually works, so….
286 00:34:08.429 ⇒ 00:34:11.699 Uttam Kumaran: I did try to implement a tool.
287 00:34:11.719 ⇒ 00:34:16.790 Bruno Vendruscolo: Lack of time. Team of one, right? So… reptile. Is this a tool?
288 00:34:16.840 ⇒ 00:34:30.849 Uttam Kumaran: Yeah, so we’re trying Reptile, Code Rabbit, Curse, but see, like, for example, there’s a task, someone on the AI team, we just got a couple new clients, so we had to push some tasks, but I said, do a spike, try all of them.
289 00:34:31.589 ⇒ 00:34:39.769 Uttam Kumaran: And then tell me which one we gotta use. But then also configure it. Like, Reptile, you can put in the instructions, right? Like, this is a dbt thing, so then…
290 00:34:39.939 ⇒ 00:34:47.569 Uttam Kumaran: So, someone needs to sort of configure and own that, but I’m like, cool, there’s CodeRabbit, Reptile, Cursor released one, Copilot has one.
291 00:34:47.869 ⇒ 00:34:50.019 Uttam Kumaran: And I said, just try all of them.
292 00:34:50.349 ⇒ 00:35:08.339 Uttam Kumaran: do a spike where we test it for some common PRs, changing a DBT model, updating Dagster, which is what we’re, like, kind of an airflow thing, right? And then see what the reviews are, customize it, and then pick. They’re all 30 to 40 bucks a user. It’s expensive, but, like, I think it’s worth it.
293 00:35:08.809 ⇒ 00:35:13.889 Uttam Kumaran: But, like, now that there’s so many options, I just don’t have the time to, like, pick, and I want to make a good decision, right?
294 00:35:14.479 ⇒ 00:35:28.249 Uttam Kumaran: So Greptile, there’s a couple there, but Greptile… I’d also, like, look, a lot of these guys, like, I was emailing with the Greptile founders, because I was asking for a startup discount. They said, yeah, we’ll give you a discount. So we can go talk to all these people who are just there if we email them, so…
295 00:35:28.449 ⇒ 00:35:37.179 Uttam Kumaran: I also think we may be the only people in data thinking about things this way. Like, I think maybe some of the bigger companies, like DBT and, like.
296 00:35:37.319 ⇒ 00:35:41.039 Uttam Kumaran: All these people with, like, unlimited money, but in consulting.
297 00:35:41.259 ⇒ 00:35:50.139 Uttam Kumaran: I don’t know if there’s gonna be other companies thinking about things like this, and I just think we’re, like, I’m handcuffed just because we’re so busy with everything else.
298 00:35:52.150 ⇒ 00:35:54.329 Bruno Vendruscolo: Yeah, I see, and this is…
299 00:35:55.100 ⇒ 00:36:04.129 Bruno Vendruscolo: This looks really great. I mean, all those initiatives, I think they all make sense. I think they bring a lot of value for the team, for clients, for the future, and how you sell.
300 00:36:04.390 ⇒ 00:36:13.060 Bruno Vendruscolo: Brain Forges, like a… accompanying the edge, like, providing the best services. I think it all makes sense, including
301 00:36:13.560 ⇒ 00:36:18.679 Bruno Vendruscolo: Hiring great talents that would work with you, ….
302 00:36:18.680 ⇒ 00:36:25.319 Uttam Kumaran: Yeah, I want people to be like… people are gonna start saying, like, if your company isn’t using AI, like, why would you want to work there, right? I don’t know.
303 00:36:25.610 ⇒ 00:36:29.370 Uttam Kumaran: Or if your company’s like, oh, I don’t know, we’re kind of worried, or like…
304 00:36:29.980 ⇒ 00:36:36.189 Uttam Kumaran: Right? I want this to be a place where people are like, yeah, they’re using everything, everywhere. It’s not like cheating. It’s like, no, no.
305 00:36:36.190 ⇒ 00:36:36.620 Bruno Vendruscolo: I think I’.
306 00:36:36.620 ⇒ 00:36:39.110 Uttam Kumaran: Mandatory, like, you have to use it.
307 00:36:39.110 ⇒ 00:36:47.740 Bruno Vendruscolo: Yeah, yeah, it’s not bullshit. I mean, we are truly using this to make things happen, to create impact, and yes, this makes sense. So….
308 00:36:47.740 ⇒ 00:37:05.879 Uttam Kumaran: But it’s not some gimmicky thing where it’s, like, using AI to write SQL, tell me how many orders I had today. Like, that’s bullshit, right? Like, it needs to get better, but it starts with these small things, but then level two, and I’ll send you, I just gave a talk about this, about how we’re thinking about, like, using AI in decision making.
309 00:37:06.080 ⇒ 00:37:22.590 Uttam Kumaran: Where, again, as I mentioned, like, our company, we have all this… we have… we’ve met with everybody and the stakeholders, we have answered all the best data questions, we have the repos, we have our slacks, and we have the data, like, the AI should be able to answer some, like.
310 00:37:22.730 ⇒ 00:37:26.820 Uttam Kumaran: Should be able to do, like, the… think about the most junior analyst.
311 00:37:27.020 ⇒ 00:37:33.450 Uttam Kumaran: it should be able to do some stuff like that, right? So, like, how far are we for that? I don’t know. I want to know.
312 00:37:33.770 ⇒ 00:37:36.109 Bruno Vendruscolo: I agree with you. I agree with you. I think…
313 00:37:36.570 ⇒ 00:37:43.370 Bruno Vendruscolo: With the right context information and setup, it should be able to do A person’s… Sure.
314 00:37:43.750 ⇒ 00:37:45.609 Uttam Kumaran: Yeah, yeah, yeah.
315 00:37:45.610 ⇒ 00:37:51.310 Bruno Vendruscolo: Got really close, … All makes sense. So….
316 00:37:51.730 ⇒ 00:37:56.650 Uttam Kumaran: So let me, let me send an email with some stuff. I’ll attach some of these things.
317 00:37:57.050 ⇒ 00:38:07.050 Uttam Kumaran: And then, yeah, I’ll get some clarification from you on, like, hours and rate and stuff like that, and then, yeah, let me, I’ll go back to the team and try to get something approved this week, and we’ll see.
318 00:38:08.090 ⇒ 00:38:12.730 Uttam Kumaran: Cool, let’s… and let’s iterate, if you want to… just to share something. Yeah.
319 00:38:13.760 ⇒ 00:38:18.619 Bruno Vendruscolo: And you get a few comments, and we start getting on track.
320 00:38:20.040 ⇒ 00:38:21.730 Bruno Vendruscolo: From scratch, like anything.
321 00:38:21.730 ⇒ 00:38:22.410 Uttam Kumaran: Oh.
322 00:38:22.410 ⇒ 00:38:24.480 Bruno Vendruscolo: Very informal. I’ll be around.
323 00:38:24.710 ⇒ 00:38:25.380 Uttam Kumaran: Okay.
324 00:38:26.050 ⇒ 00:38:27.650 Uttam Kumaran: Okay, dude, great.
325 00:38:28.000 ⇒ 00:38:35.460 Bruno Vendruscolo: Well, thank you so much for the time, Dimilare. Nice chatting with you. Thank you for promoting me.
326 00:38:37.620 ⇒ 00:38:44.849 Bruno Vendruscolo: Always, always great talking to you. You too, Tom. Thank you so much again for… for the time.
327 00:38:44.850 ⇒ 00:38:48.510 Uttam Kumaran: If that can be helpful in any other way, please let me know. Happy to.
328 00:38:48.690 ⇒ 00:38:49.340 Uttam Kumaran: And, ….
329 00:38:49.340 ⇒ 00:38:49.940 Bruno Vendruscolo: Sure.
330 00:38:49.940 ⇒ 00:38:51.909 Uttam Kumaran: Yeah, of course. Okay.
331 00:38:52.110 ⇒ 00:38:52.770 Bruno Vendruscolo: Sure.
332 00:38:53.430 ⇒ 00:38:54.650 Bruno Vendruscolo: Thank you, Ben.
333 00:38:54.650 ⇒ 00:38:57.270 Uttam Kumaran: Alright, thank you guys, appreciate it, have a great day.
334 00:38:58.010 ⇒ 00:38:59.510 Uttam Kumaran: You too. Bye-bye.