Meeting Title: Brainforge AI Projects Check-in Date: 2025-12-04 Meeting participants: Pranav, Uttam Kumaran
WEBVTT
1 00:02:39.080 ⇒ 00:02:40.500 Uttam Kumaran: Hey, man, can you hear me?
2 00:02:42.960 ⇒ 00:02:44.750 Pranav: Yay! Yeah, I can hear you, can you hear me?
3 00:02:46.080 ⇒ 00:02:48.100 Uttam Kumaran: Sorry, one second, can you hear me now?
4 00:02:48.500 ⇒ 00:02:49.749 Pranav: Yeah, I can hear you. Can you hear me?
5 00:02:49.800 ⇒ 00:02:51.569 Uttam Kumaran: Yes, yes, how are you?
6 00:02:51.800 ⇒ 00:02:53.120 Pranav: Good, good, how are you?
7 00:02:53.120 ⇒ 00:02:55.180 Uttam Kumaran: Good, dude. Busy.
8 00:02:55.180 ⇒ 00:02:56.180 Pranav: Yeah, yeah.
9 00:02:56.320 ⇒ 00:02:58.399 Pranav: Sounds some good news, that’s great.
10 00:02:58.400 ⇒ 00:03:02.299 Uttam Kumaran: Yeah, so we’ve just been growing a bunch last,
11 00:03:02.540 ⇒ 00:03:18.640 Uttam Kumaran: last, like, few months, like, last three months especially, so… Yeah, we closed probably, like, another two, like, more AI application-related clients, so… thought I’d give you a call, see if, see what you’re up to, and see if you’d still be interested in doing stuff.
12 00:03:19.280 ⇒ 00:03:23.610 Pranav: That’s awesome, yeah, yeah, I’m, it’s a good timing. I just closed with, like.
13 00:03:23.740 ⇒ 00:03:32.030 Pranav: two clients, like, last week, so I, I have a lot of room in my calendar. So, yeah, I’m glad this worked out.
14 00:03:32.640 ⇒ 00:03:38.219 Uttam Kumaran: Cool. What’s the… what’s, like, what have you been up to, and, like… Like, what’s… what’s stayed in your world?
15 00:03:38.860 ⇒ 00:03:45.279 Pranav: Yeah, so, more or less the same stuff that we were talking about, like, a few months ago.
16 00:03:45.760 ⇒ 00:03:55.609 Pranav: Basically… Building AI automations, and I picked up a client, in, like, Austin.
17 00:03:55.890 ⇒ 00:04:11.279 Pranav: I think it was around August, and been working with them part-time, continuing to work with them part-time, to basically build, like, no-code, like, automations for their provider ops team, or for their operations team.
18 00:04:11.450 ⇒ 00:04:12.969 Pranav: And,
19 00:04:13.200 ⇒ 00:04:21.000 Pranav: Yeah, so that… I think, actually, I reached out to you, like, a while ago, kind of talking about, like, Make versus N8N.
20 00:04:21.000 ⇒ 00:04:21.890 Uttam Kumaran: Yeah, yeah.
21 00:04:21.899 ⇒ 00:04:24.449 Pranav: gum loop, and so that was for this client, because.
22 00:04:24.450 ⇒ 00:04:25.030 Uttam Kumaran: Okay.
23 00:04:25.030 ⇒ 00:04:33.260 Pranav: They had not had a ton of, experience using Make specifically, and they were ready to close with Make because of,
24 00:04:33.420 ⇒ 00:04:35.240 Pranav: certain,
25 00:04:35.840 ⇒ 00:04:54.749 Pranav: certain, like, security reasons, as well as just the pricing was right. I had built some stuff with N8in just, like, for personal projects, but now, just kind of using these applications, like, at scale, integrating them with, like, a bunch of different applications, that’s been…
26 00:04:54.770 ⇒ 00:04:56.949 Pranav: what I’ve been doing for the past…
27 00:04:57.060 ⇒ 00:05:02.030 Pranav: What, August was, what, almost 4 months ago at this point?
28 00:05:02.030 ⇒ 00:05:02.990 Uttam Kumaran: Nice.
29 00:05:02.990 ⇒ 00:05:13.519 Pranav: Yeah, and so that was for one client, and then for everything else, basically, just, like, software engineering, what I’ve been used to for the last, like, 4 years, and…
30 00:05:13.600 ⇒ 00:05:28.780 Pranav: just leveraging, like, AI tools. So, like, getting more used to, like, Cloud code, using, like, like, building, like, custom agents with, like, CloudMD files, like, things of that nature. So, yeah, just a little bit of both, I would say.
31 00:05:29.170 ⇒ 00:05:36.029 Uttam Kumaran: Okay, dope. Yeah, we’re still doing a lot of, like, agent building. We’re not doing as much…
32 00:05:36.150 ⇒ 00:05:44.629 Uttam Kumaran: Like, NA then, I would say. We’ve actually been using Mastra, which is, like, a JavaScript, like…
33 00:05:44.790 ⇒ 00:05:51.840 Uttam Kumaran: agent framework, and moving, kind of, like, honestly, moving a lot off of N8N, just because, like.
34 00:05:52.110 ⇒ 00:06:06.869 Uttam Kumaran: I don’t know, we can’t, like, do a lot of prompt management, and some of this stuff is just getting more sophisticated, so we’re kind of moving everything to code, which has been good, it’s been fun. And so, yeah, we’re building some… we’re still
35 00:06:06.920 ⇒ 00:06:16.580 Uttam Kumaran: building and maintaining sub-agents, still doing a lot of, like, internal agent development. And then, yeah, we just… now we have two…
36 00:06:17.100 ⇒ 00:06:20.939 Uttam Kumaran: two clients where we’re building, like, AI applications for, and then, like.
37 00:06:21.080 ⇒ 00:06:24.129 Uttam Kumaran: Kind of, like, several more in the pipeline, so it’s been good.
38 00:06:25.120 ⇒ 00:06:27.239 Pranav: That’s awesome. Yeah, yeah,
39 00:06:27.400 ⇒ 00:06:37.030 Pranav: you… are you seeing, like, there’s essentially just more custom… customizable, like, areas, I guess, when you’re using, like, code versus N8N?
40 00:06:37.730 ⇒ 00:06:46.859 Uttam Kumaran: Yeah, it’s just like… I don’t know, I just feel like it’s more reliable, like, we’ve had some workloads on NAN, they just crashed because we, like, exhausted memory.
41 00:06:46.930 ⇒ 00:06:49.280 Pranav: And…
42 00:06:49.280 ⇒ 00:07:00.089 Uttam Kumaran: I need, like… I start asking for stuff like staging and, like, CICD pipelines, and wanting to do unit testing, and a lot of that is just hard to do in any of them.
43 00:07:00.090 ⇒ 00:07:02.469 Pranav: Totally. And some of the stuff we’re doing for clients.
44 00:07:02.510 ⇒ 00:07:03.510 Uttam Kumaran: It’s just…
45 00:07:03.670 ⇒ 00:07:16.509 Uttam Kumaran: way more serious now than when I was just, like, messing around with NNN stuff. So I feel like we’re still, like, probably prototyping in NNN, but frankly, now that we even have, like.
46 00:07:16.570 ⇒ 00:07:27.940 Uttam Kumaran: everybody’s using Cursor, it’s getting so much quicker to just, like, develop everything as code, and so we’re actually migrating a lot of our stuff off of N8N onto our own.
47 00:07:28.070 ⇒ 00:07:44.629 Uttam Kumaran: like, we’re hosting some stuff internally on Heroku, but, you know, whatever. And so, for the most part, we’re able to sort of build, like, JS apps, primarily using this master framework, which has been great. Yeah.
48 00:07:44.630 ⇒ 00:07:45.210 Pranav: God.
49 00:07:46.680 ⇒ 00:07:53.749 Uttam Kumaran: But, I mean, it’s like, it’s a mix of, like, it’s a mix of still doing, like, normal software, like, full-stack development, and then…
50 00:07:53.950 ⇒ 00:07:58.449 Uttam Kumaran: then it’s like, we all kind of had to learn this, like, kind of AI framework, so…
51 00:08:00.160 ⇒ 00:08:02.020 Pranav: Yeah, totally. So…
52 00:08:02.360 ⇒ 00:08:16.180 Pranav: have you… because I would say this too, like, with building with code, and maybe going off of, like, any den, people may think that, like, the deployment time is, like, a lot longer, but I haven’t really…
53 00:08:16.460 ⇒ 00:08:28.729 Pranav: felt that as… it hasn’t been as difficult, I would say. I think still going 0 to 1 with, like, all these, like, tools, like you said, like cursor, or, like, Claude Code, like, it’s still pretty quick.
54 00:08:28.730 ⇒ 00:08:41.049 Pranav: And I would say my experience with using, like, these no-code tools, or, like, low-code tools, is that I’m not really building applications that need to be…
55 00:08:41.169 ⇒ 00:08:43.309 Pranav: Worrying about memory, and… Yeah.
56 00:08:43.919 ⇒ 00:08:53.679 Pranav: Like, that, like, one… one, like, module-fits-all type of environment, I think works well for, like.
57 00:08:53.939 ⇒ 00:08:56.649 Pranav: Really simple agents, or just, like…
58 00:08:56.840 ⇒ 00:08:57.400 Uttam Kumaran: Yeah.
59 00:08:57.400 ⇒ 00:09:03.640 Pranav: that you were building in Zapier before, and then just require maybe… Exactly. Yeah.
60 00:09:04.000 ⇒ 00:09:21.470 Uttam Kumaran: Yeah, and that’s kind of how I felt, but it also took, like, it took a lot of us to upskill a little bit, and… and also, like, now we’re… I’m trying to charge way more and build, like, a real sophisticated product, so… for us, it’s like, okay, we’re just gonna build actual
61 00:09:21.770 ⇒ 00:09:31.720 Uttam Kumaran: you know, product. And, like, I never was worried about that. It was more just, like, there wasn’t a great framework in JavaScript until we found this one.
62 00:09:31.860 ⇒ 00:09:41.909 Uttam Kumaran: And it’s been great to sort of start to move stuff, and then just everything gets better, like, scalability, and then also just, like, for me, the most important thing is, like.
63 00:09:42.020 ⇒ 00:09:58.060 Uttam Kumaran: being able to do observability and prompt management and, like, error logging, and not having to, like… like, for N8N, you need, like, Enterprise to get a bunch of logs out. You… it’s, like… I… I just, like, we’re trying to build, like, real solutions, and I can go charge, you know, like.
64 00:09:58.060 ⇒ 00:10:12.820 Uttam Kumaran: way more for it, you know? And so… and then, for me, it’s also like, okay, if we figured out, like, how to get it closer to real engineering, then when I go, you know, bring people on, I can bring in people with, like, full-stack background, they just have to figure out the AI part.
65 00:10:12.820 ⇒ 00:10:20.089 Uttam Kumaran: Versus also go figure out, like, N8N. And so I think we still have some clients that are asking us for that, but frankly, I think
66 00:10:20.190 ⇒ 00:10:32.240 Uttam Kumaran: a lot of them, like, if they’re asking for N8N, it’s clear there’s probably nobody in the company that’s, like, very technical, and yeah, they’re just looking for, like, Zapier Plus versus, like…
67 00:10:32.440 ⇒ 00:10:48.690 Uttam Kumaran: sort of a real, like, more sophisticated solution, which… which I get, like, that’s who we were going after before, and now I think we’re just… we’re just going after bigger… bigger fish. We’re just, like, building more applications, and a lot of the, like, SDKs and integrations we want to use, we couldn’t…
68 00:10:48.750 ⇒ 00:10:53.690 Uttam Kumaran: we couldn’t use in, in ANN, like, we’d have to…
69 00:10:53.790 ⇒ 00:11:00.309 Uttam Kumaran: build, like, a webhook, or, like, call some REST API, versus we can now… if it’s all as code, it’s… it’s chill.
70 00:11:00.750 ⇒ 00:11:02.170 Pranav: Right, right, right.
71 00:11:02.590 ⇒ 00:11:03.200 Uttam Kumaran: Yeah.
72 00:11:04.640 ⇒ 00:11:18.250 Uttam Kumaran: Yeah, let me tell you about this project. So, we, we signed with this company called Lilo Social. They’re, like, an e-commerce agency. They make a lot of money managing, you know, ad budgets for a bunch of different clients.
73 00:11:18.280 ⇒ 00:11:30.709 Uttam Kumaran: And we’re helping them build, like, a AI-powered, like, forecasting tool that they can use internally to, like, forecast ad budgets, and then also building, like, a module for them that’s helping them, like.
74 00:11:31.920 ⇒ 00:11:46.199 Uttam Kumaran: figure out, like, a new creative, for, like, ads to test. And they actually were using… they were actually working with another firm that kind of just, like, sucks. And then it’s like a…
75 00:11:46.220 ⇒ 00:11:58.139 Uttam Kumaran: the CEO of this company is, like, friends of a friend of mine, and was like, hey, this other agency, like, they’re kind of gatekeeping stuff, there’s a lot of bug, they’ve been taking a long time, and, like, we just need, like, help.
76 00:11:58.250 ⇒ 00:12:06.639 Uttam Kumaran: And then we looked through a bunch of their stuff, and so right now, we signed for, like, a 3-month contract. We have, like, about a month or so to just, like.
77 00:12:06.820 ⇒ 00:12:17.250 Uttam Kumaran: wrangle all the existing code, like, build, like, a monorepo out of everything. They have, like, a bunch of scattered repos, and then host it on…
78 00:12:17.610 ⇒ 00:12:24.159 Uttam Kumaran: this company’s infrastructure, and then also rebuilt Auth. The other comp… the company that they were…
79 00:12:24.340 ⇒ 00:12:26.970 Uttam Kumaran: Using before, they had, like, their own, like, auth.
80 00:12:27.180 ⇒ 00:12:33.579 Uttam Kumaran: repo that they were using that’s sort of, like, homegrown, and it was, like, whatever. So, we have, like, about a month of just, like.
81 00:12:33.700 ⇒ 00:12:46.929 Uttam Kumaran: getting the codebase into a place where it’s, like, running on their own infra, and, like, the most of, like, the basic bugs are fixed, and then we have about… we have another two months to sort of build those two modules, which I feel like is…
82 00:12:47.130 ⇒ 00:12:50.519 Uttam Kumaran: for the most part, it’s actually, like, plenty of time.
83 00:12:51.100 ⇒ 00:12:55.690 Uttam Kumaran: And so right now, it’s just, like, I’m… I’m sort of just, like…
84 00:12:55.710 ⇒ 00:13:15.270 Uttam Kumaran: kind of there on the project, like, doing a little bit of scoping, but there’s one other guy, his name is Cerf, and he’s… he’s sort of, like, kind of, like, architect level, and so, yeah, I kind of thought, like, you know, if you have time, there’s probably at least, like, 10 to 20 hours of work,
85 00:13:15.270 ⇒ 00:13:29.039 Uttam Kumaran: you know, just for that client a week. I know, of course, like, you can come take a look and let me know what you think. And then we’re also, like, we have some… it would be great if you’re on that client, you’ll also start to see some of the
86 00:13:29.040 ⇒ 00:13:40.849 Uttam Kumaran: work we’re trying to do internally, and then we have another… we have, like, one other client as well that you can have a look at. So I feel like we got… we have, like, you know, at least 20 hours a week of work.
87 00:13:40.860 ⇒ 00:13:46.349 Uttam Kumaran: You know, if you’re interested. And then our team right now, it’s like, we actually have…
88 00:13:46.360 ⇒ 00:14:03.409 Uttam Kumaran: two sort of, like, mid-level, like, engineers that are kind of, like, full-stack. We have one, like, sort of, like, much more, much more senior. He sort of, like, leads AI from the technical side. Surf, who’s honestly more of, like, a, like, full-stack.
89 00:14:03.410 ⇒ 00:14:14.150 Uttam Kumaran: But, like, it’s starting to take on… like, he’s just really good at building, you know, product, and we have one product manager that’s helping out internally. So, like, pretty good team, everybody’s, like, really, really chill, so…
90 00:14:15.190 ⇒ 00:14:25.909 Pranav: Yeah, no, that’s, that’s awesome. There was a few things I really liked, like the fact that you said there is somebody that, you know, there are a few mid-level engineers, but there’s also some people that are, like.
91 00:14:26.490 ⇒ 00:14:42.909 Pranav: that we feel can be, like, a little bit more of, like, that senior level. Yeah. I think that’s… that’s good. I’m at a point where I would say, like, I’m mid-level, trying to get to that point where maybe I would say, like, a lot of companies may hire me as a senior.
92 00:14:43.140 ⇒ 00:14:43.770 Uttam Kumaran: Yeah.
93 00:14:43.770 ⇒ 00:14:50.349 Pranav: But I think the senior kind of title is getting just thrown around a lot at, like, smaller companies.
94 00:14:50.550 ⇒ 00:14:59.759 Pranav: And so, when I think of, like, becoming a senior engineer, I think about, like, at, like, these, like, fan companies, like, where senior engineers maybe
95 00:14:59.860 ⇒ 00:15:07.539 Pranav: thought of to actually, like, satisfy, like, specific requirements, like, really understanding of system design, like, application design, things.
96 00:15:07.540 ⇒ 00:15:08.050 Uttam Kumaran: Yeah.
97 00:15:08.050 ⇒ 00:15:16.369 Pranav: That’s where I definitely want to, like, improve in a little bit. I think that’s, like, the next step for me, because I feel pretty confident in my…
98 00:15:16.670 ⇒ 00:15:33.040 Pranav: you know, being a mid-level engineer and, like, performing on, like, a design made by somebody else, and also creating some designs and, like, being part of, like, that conversation, but just not fully owning it. I think… I’m just happy that I wouldn’t be thrown into that situation.
99 00:15:33.040 ⇒ 00:15:37.329 Uttam Kumaran: No, and like, my company before, it was… it was…
100 00:15:37.380 ⇒ 00:15:42.159 Uttam Kumaran: It was a lot less of, like… this was, like, a lot safer of a situation.
101 00:15:42.160 ⇒ 00:15:58.179 Uttam Kumaran: Because, one, like, I was always there, but, like, I think now what I… what we kind of… we have… we have a guy named Sam, he’s great. He used to be, like, a former CTO of a couple startups, really great. And then, yeah, he’s sort of guiding the rest of the team on, like.
102 00:15:58.180 ⇒ 00:16:04.659 Uttam Kumaran: okay, how to actually build fundamentally, like, great… how to ship great code. We do a lot of, like, code reviews, spikes.
103 00:16:04.680 ⇒ 00:16:15.850 Uttam Kumaran: like, architecture reviews, so, like, I love that because that’s how I used to run, like, my teams when I was in… when I was, you know, leading product. So, we’re doing a lot of work like you would expect.
104 00:16:15.850 ⇒ 00:16:31.480 Uttam Kumaran: like, a fast-growing, sort of, engineering, like, firm is doing. Not, like, a startup where, like, we’re just shipping, like, crap. It’s actually… we’re trying to do more spikes around, like, for example, like, okay, what’s the best, like, text-to-SQL framework? Okay, let’s, like, run a spike
105 00:16:31.480 ⇒ 00:16:49.570 Uttam Kumaran: test a bunch of things, make a recommendation, and then, yeah, we do, like, architecture reviews, code reviews. So yeah, it’s fun. I think… I think still the… there’s, like, pressure, because everybody I mentioned is working on at least, like, 2 or 3 clients. So it’s not like…
106 00:16:49.680 ⇒ 00:17:00.500 Uttam Kumaran: it’s not, like, completely… everybody’s still working pretty hard, but it’s actually a lot safer because we have more redundancy. Like, on every client, there’s at least 2 or 3 people in a
107 00:17:00.830 ⇒ 00:17:16.639 Uttam Kumaran: And that way, it’s like, if you’re out, you’re not, like, it’s not like end of the world, and if things break, there’s other people. So we’re usually staffing every client with at least three people. The way it’s kind of working is, like, when we get a new client like Lilo, for example, I start, or me and one people, one person starts.
108 00:17:16.640 ⇒ 00:17:23.510 Uttam Kumaran: we sort of figure it out, like, after a week or two, and then we start to loop people in. Versus before, I used to just sort of, like.
109 00:17:23.510 ⇒ 00:17:35.020 Uttam Kumaran: okay, let’s all figure it out, and then it would be, like, too many people, some people are coming in and out, and so right now, what we’re doing is, like, we’ll kick off a project, be, like, one or two weeks of, like, discovery, and then I’ll slowly loop in
110 00:17:35.020 ⇒ 00:17:48.500 Uttam Kumaran: whether it’s, like, someone from our data team, from our, like, analyst team, or, like, on engineering. And so it’s usually, like, me and Sam, or me and a senior person, and then I’m sort of like, okay, here’s the scope. I’ll still run, like, point on…
111 00:17:48.520 ⇒ 00:17:59.980 Uttam Kumaran: like, client meetings, basically, because that could just be stress. But, like, I’m down if people… if people want to do that, also. So that’s why I think there’s fun opportunity here, like, you can actually…
112 00:18:00.010 ⇒ 00:18:11.149 Uttam Kumaran: like, if you want it, you could actually definitely direct, like, you know, a lot of… make a lot of, like, sort of product manager type decisions as well, but just on the code side, yeah, I mean, we’re doing, like.
113 00:18:11.300 ⇒ 00:18:17.949 Uttam Kumaran: pretty significant stuff on the AI side, and it’s just getting more and more complex and more interesting, so…
114 00:18:18.840 ⇒ 00:18:32.409 Pranav: That’s awesome. Yeah, it sounds… it sounds really good to me. Like, I like the fact that, because I’ve been on projects before where it was just me as a developer, and then maybe, like, a very much, like, part-time, like, PM.
115 00:18:32.770 ⇒ 00:18:33.600 Uttam Kumaran: Yeah.
116 00:18:33.600 ⇒ 00:18:37.779 Pranav: it’s, when things are breaking, it’s a little bit difficult, I would say, because…
117 00:18:37.780 ⇒ 00:18:38.180 Uttam Kumaran: Yes.
118 00:18:38.180 ⇒ 00:18:39.489 Pranav: There’s a lot of…
119 00:18:39.490 ⇒ 00:18:42.129 Uttam Kumaran: And they’re like, I can’t help you, you know?
120 00:18:42.130 ⇒ 00:18:53.659 Pranav: Yeah, yeah, yeah, and if you… I feel like when engineers are able to just, like, bounce things off of each other, it’s more than just, like, two engineers, it feels more like a team. And then when you have three, it just, like…
121 00:18:53.950 ⇒ 00:19:07.670 Pranav: it’s, a lot more… it’s not just, like, 3 times the productivity most of the time. And I feel like that it’s especially apparent when, like, there are bugs or, maybe even just, like, design decisions.
122 00:19:07.670 ⇒ 00:19:14.929 Uttam Kumaran: Yeah, and that’s exactly what changed when we… when we got to, like, finally, like, 3 people, and then… and… and me.
123 00:19:15.300 ⇒ 00:19:34.369 Uttam Kumaran: Which was, like, the three of them can now meet, and, like, now it’s, like, sort of a really good pod, you know, where, like, we meet, we all meet in the morning, but then, like, we talk, and then it’s, like, AI team goes and meets on priorities, and, like, yeah, it’s actually a really good squad, like, I think you’ll like everybody. And then, yeah, I mean, dude, like, I want you to suss it out, and then…
124 00:19:34.370 ⇒ 00:19:53.449 Uttam Kumaran: like, if you end up, like, hey, like, this is great, like, I feel good with the team, like, we’re… we’re gonna close another, like, 2 or 3 clients, and, like, you have a place here if it all works out. So, I mean, that’s what, ultimately, like, I’d be looking for, which is, like, I know we’re just saying, like, hey, if you want to come work 10, 20 hours, but for me.
125 00:19:53.450 ⇒ 00:20:07.729 Uttam Kumaran: everybody that’s, like, on the team now, even part-time, it’s always like, okay, if things work out, like, how do we get, you know, all of your time so that we can… for me, it’s like, I can go just sell bigger, more complicated stuff, so that would be, like, our path.
126 00:20:08.540 ⇒ 00:20:18.230 Pranav: Yeah, that’s, that’s definitely, like, the way I’m trying to move towards. I’d rather just have all my investment with just, like, one company.
127 00:20:18.230 ⇒ 00:20:36.220 Uttam Kumaran: Yeah, and I remember we talked about the other agency, too, that you were working for, and I feel like we were probably similar to them when we were smaller, but now it’s, like, we’re just much, much bigger, and, like, it’s getting a lot more, like, formalized, but less… but still, like, we’re doing really…
128 00:20:36.390 ⇒ 00:20:54.250 Uttam Kumaran: sophisticated work at a really good rate, but it’s… it’s, like, it’s not crazy, and we’re also not… we’re just… it’s just getting better, like, even… I’m doing a lot of… we’re doing a lot of cool work that, like, I want to sell, which is, like, real sophisticated AI, you know, agent building, and things beyond just, like, N8N and stuff like that.
129 00:20:54.940 ⇒ 00:20:56.649 Pranav: Yeah, totally, so…
130 00:20:57.000 ⇒ 00:21:12.850 Pranav: Yeah, with, like, that previous, like, agency, too, like, I think when you’re starting up, there’s not gonna be as many engineers, and maybe they had more clients that they could staff people on, so, like, yeah, I did have, like, a few clients where I was, like, the sole engineer. And you learn a lot that way, too.
131 00:21:12.850 ⇒ 00:21:21.589 Pranav: And so, but I would say, like, yeah, it sounds like the point that you guys are at right now is just more, interesting to me, because I feel like…
132 00:21:21.590 ⇒ 00:21:31.780 Uttam Kumaran: selling deals where we can staff them, like, appropriately, which wasn’t always the case. Now, it’s like, I won’t do… like, we now have minimums, like, because I can’t guarantee
133 00:21:31.860 ⇒ 00:21:34.760 Uttam Kumaran: Not only success, but that my team will be…
134 00:21:34.830 ⇒ 00:21:40.210 Uttam Kumaran: like, happy with me. So… which, like, really, really matters. And…
135 00:21:40.210 ⇒ 00:21:56.009 Uttam Kumaran: for me, it’s also just, I want to get cool work, like, I want to go do, like, image generation, I want to go do, like, we started doing some voice work, like, that’s the sort of stuff that’s still really fun. And then, also, like, everyone spends, like, 10 or 20% of their time working on internal projects.
136 00:21:56.110 ⇒ 00:22:03.989 Uttam Kumaran: Which is great, because we’ve built stuff for marketing, for sales, and, like, everybody in our company uses AI, like, non-stop.
137 00:22:04.780 ⇒ 00:22:11.739 Pranav: Yeah, that’s awesome, and then you can have, like, all hands on deck for, like, those internal things. People are, like, less siloed.
138 00:22:11.740 ⇒ 00:22:15.960 Uttam Kumaran: That’s also, like, it’s, like, sort of budgeting out, like, 10-20% of time to just, like.
139 00:22:16.080 ⇒ 00:22:29.629 Uttam Kumaran: do internal stuff, which, like, we built voice agents internally, and, like, we just did things that, like, no client… like, I asked us for, and it’s, like, more complicated than what any client has, like, asked us for before, so…
140 00:22:30.360 ⇒ 00:22:33.730 Pranav: Right, yeah, that makes sense, and it’s like…
141 00:22:34.480 ⇒ 00:22:39.200 Pranav: it’s, those are all, you know, hours that aren’t billable, right? So…
142 00:22:39.200 ⇒ 00:22:55.300 Uttam Kumaran: Yeah, but that’s also, like, I’m paying… I always… I’m paying out of that, but I know the ROI for the company, and, like, ultimately, like, yeah, but you don’t build any camaraderie with your team if you’re just, like, on one client, and there’s… it’s, like, not a company, you know?
143 00:22:55.520 ⇒ 00:23:14.920 Uttam Kumaran: And so… so… but early on, I, like, couldn’t afford to do that, and so now… but it’s always been, like, I’ve always been working on stuff internally for projects, and now I’m just happy we can at least afford, like, 10 or 20% of the time, and otherwise you just kind of sit there, and you don’t talk to anyone, like, I can’t deal with that, you know, so…
144 00:23:14.920 ⇒ 00:23:20.659 Pranav: Yeah, yeah, yeah. No, I felt like from that first conversation that we had, I don’t…
145 00:23:20.780 ⇒ 00:23:23.070 Pranav: I don’t even remember exactly when that was.
146 00:23:23.070 ⇒ 00:23:26.260 Uttam Kumaran: Yeah, I don’t know, maybe, like, 6 months ago or something, but…
147 00:23:26.260 ⇒ 00:23:27.670 Pranav: Has it already been 6 months?
148 00:23:27.670 ⇒ 00:23:33.100 Uttam Kumaran: I feel like it was a while ago, but I don’t know, my time is kind of a little bit warped, so I’m not sure, actually.
149 00:23:33.310 ⇒ 00:23:45.370 Pranav: Yeah, I… it… I was gonna say 3 months ago, but I think maybe it was, like, 4… like, 4 to 6 months ago, yeah. I did… I did like a lot of the things that you were talking about, and it sounds like you guys are…
150 00:23:45.920 ⇒ 00:23:58.270 Pranav: kind of executing, pretty well there as well. And so, yeah, it sounds like it aligns to me. I… like you said, like, maybe I can talk to, like, the team,
151 00:23:58.270 ⇒ 00:23:58.750 Uttam Kumaran: Yeah.
152 00:23:58.750 ⇒ 00:23:59.250 Pranav: Maybe look…
153 00:23:59.250 ⇒ 00:24:08.019 Uttam Kumaran: Last time you ended up talking to Sam on the team, but yeah, I’ll connect you again. And, I invited Serf, but I think he got dragged into another meeting.
154 00:24:08.020 ⇒ 00:24:10.010 Pranav: So I’ll connect you with…
155 00:24:10.010 ⇒ 00:24:28.379 Uttam Kumaran: with both of them, and then, yeah, Surf is the primary on Lilo, so he’ll kind of tell you what he’s thinking. And then, yeah, Surf works on two clients for us, and then, I’ll have you chat with Sam, and then, yeah, if you feel good about it, like, we can, you know, onboard as soon as next week.
156 00:24:29.140 ⇒ 00:24:31.449 Pranav: Oh, that would be great, yeah.
157 00:24:31.800 ⇒ 00:24:38.729 Pranav: So, I am going to be a little bit spotty for tomorrow and Monday. Okay.
158 00:24:38.840 ⇒ 00:24:40.270 Pranav: But… Yeah, it’ll…
159 00:24:40.270 ⇒ 00:24:47.720 Uttam Kumaran: Probably be Monday, although I can… I’ll send the email. If Surf can meet, then yeah, you can coordinate with him.
160 00:24:47.940 ⇒ 00:24:50.020 Uttam Kumaran: Yeah, it’s Friday, so I don’t know.
161 00:24:51.160 ⇒ 00:24:53.499 Uttam Kumaran: Or, yeah, it’s Friday, right? Yeah, yeah.
162 00:24:53.500 ⇒ 00:24:54.710 Pranav: Today’s Thursday, but…
163 00:24:54.710 ⇒ 00:24:55.590 Uttam Kumaran: Yeah, yeah, yeah, yeah.
164 00:24:55.870 ⇒ 00:25:04.239 Pranav: Yeah, so, but yeah, tomorrow, if I’m… I’ll try to set something up with surf, if not, I can probably make Monday, like, later in the day work. Okay.
165 00:25:04.450 ⇒ 00:25:07.660 Pranav: Yeah, the sooner the better, so… Okay.
166 00:25:07.660 ⇒ 00:25:12.209 Uttam Kumaran: So, yeah, worst case, I may just, like, text you and surf.
167 00:25:12.310 ⇒ 00:25:31.260 Uttam Kumaran: Because he just does better over text, so I’ll… I’ll call him and see what he says, and then, yeah, I’ll try to connect you guys. He’s a good dude. So just… I would just ask him about, like, the project and stuff like that, and then I told him that, like, hey, if we get the deal, like, we do a couple weeks of discovery and it looks good, like, I’ll call you again, so…
168 00:25:31.310 ⇒ 00:25:33.270 Uttam Kumaran: Yeah, that’s perfect.
169 00:25:33.740 ⇒ 00:25:38.200 Pranav: That’s awesome. Yeah, yeah, text works for me as well, I’ll be able to respond probably fastest there.
170 00:25:38.200 ⇒ 00:25:39.889 Uttam Kumaran: Okay, okay, sick.
171 00:25:41.130 ⇒ 00:25:44.090 Pranav: Cool. Alright, well, this is exciting. I’m glad we’re, getting.
172 00:25:44.090 ⇒ 00:25:54.769 Uttam Kumaran: No, no, it’s been a while, I’m trying, I’m trying my best, but we, dude, like, in the last 3 months, we have, like, almost doubled the company in terms of revenue. Like, we’re just…
173 00:25:54.890 ⇒ 00:26:00.840 Uttam Kumaran: Starting to scale up, and for me, like, the biggest thing is that we can’t, like.
174 00:26:01.020 ⇒ 00:26:11.920 Uttam Kumaran: we just can’t let clients down, and we have to keep, like, executing, and then also for me to go, like, I want to go sell, like, real tough AI work for, like, 100 grand a month.
175 00:26:12.060 ⇒ 00:26:19.189 Uttam Kumaran: And I need, like, really great people who can, like, manage their shit and, like… crush it, and so…
176 00:26:19.250 ⇒ 00:26:31.850 Uttam Kumaran: like, I’m never… I know, I was like, okay, even when we were talking, I was like, yeah, I don’t know, we just didn’t have the cash, and we didn’t have the projects, but we’re sort of getting to a really good point now where I’m able to get great rates, and…
177 00:26:31.850 ⇒ 00:26:40.329 Uttam Kumaran: People are just really impressed with how fast we work, and how all of our team is, like, super, super reasonable, like, great people, so… yeah, it’s been fun.
178 00:26:40.850 ⇒ 00:26:45.509 Pranav: That’s awesome. Yeah, for me, like, environment’s huge, you know? So…
179 00:26:45.510 ⇒ 00:27:03.909 Pranav: that’s what always, like, stuck out before, like, when we were talking, like, it sounds like you’re trying to build, like, a really cool environment. Just, like, just from talking to you, like, I know that that’s going to be the case. And so, yeah, I’m excited, like, I’m looking forward to, like, you know, getting in contact with Surf, and then, yeah, if we can get started, like, sometime next week, like, that’s… that sounds amazing.
180 00:27:04.450 ⇒ 00:27:09.039 Uttam Kumaran: Cool, dude. Okay, alright, let me send a text over, and yeah, you’ll hear from me in a sec.
181 00:27:09.700 ⇒ 00:27:11.570 Pranav: Perfect, perfect. Thanks so much.
182 00:27:11.570 ⇒ 00:27:13.470 Uttam Kumaran: Okay, thank you, dude. I’ll talk to you soon.
183 00:27:13.670 ⇒ 00:27:14.359 Pranav: See ya.