Meeting Title: Brainforge x Keith Saver Intro Call Date: 2025-08-14 Meeting participants: Keith Saver, Uttam Kumaran
WEBVTT
1 00:06:07.660 ⇒ 00:06:08.870 Uttam Kumaran: Hi, Keith.
2 00:06:09.110 ⇒ 00:06:10.249 Keith Saver: Hey, how’s it going?
3 00:06:10.540 ⇒ 00:06:26.660 Uttam Kumaran: Hey, good. Sorry, I just, … this is just a crazy week. We are growing a lot. That’s exciting, yeah. We are growing a lot, but, you know, it comes with challenges, so… I’m so sorry for the delay in moving things around.
4 00:06:26.920 ⇒ 00:06:29.249 Keith Saver: No, no, no, no worries, totally get it. How’s it going?
5 00:06:29.250 ⇒ 00:06:36.899 Uttam Kumaran: I was good. I was just gonna run to go get a coffee, I’ve just been sitting at my desk for, like, 5 hours. If you don’t mind, I’ll just be off, off video.
6 00:06:36.900 ⇒ 00:06:38.799 Keith Saver: Yeah, yeah, yeah, no, that’s… that is totally fine.
7 00:06:39.050 ⇒ 00:06:40.420 Uttam Kumaran: Okay, thank you, thank you.
8 00:06:40.420 ⇒ 00:06:43.239 Keith Saver: I’m like, yeah, yeah, yeah. No, I totally get it.
9 00:06:44.060 ⇒ 00:06:45.050 Uttam Kumaran: How’s everything?
10 00:06:45.310 ⇒ 00:06:50.499 Keith Saver: Good, good! I’m doing well. I, … yeah, I don’t know how much you heard from Jamie, so I’m in my sort of.
11 00:06:50.500 ⇒ 00:07:02.079 Uttam Kumaran: I heard a little bit, but yeah, I am interested in sort of the story, and you know, you’re doing what we all wish we could do in data, is do something more notable, I think, but I would love to hear it.
12 00:07:02.080 ⇒ 00:07:08.639 Keith Saver: Yeah, no, absolutely. Yeah, so I mean, I’ve worked in, you know, a variety of data roles for, I guess, almost a decade. …
13 00:07:08.840 ⇒ 00:07:14.030 Keith Saver: Yeah, you know, most of my career has actually sort of more been in, like, nonprofits and health tech.
14 00:07:14.080 ⇒ 00:07:32.429 Keith Saver: I see, okay. A lot of healthcare stuff, and then, yeah, most recently was at Shopify, which is sort of a, you know, divergence from that. I’ve also mostly worked at startups, so obviously, you know, Shopify was kind of a… part of the reason I actually wanted to go to Shopify is I was like, you know, I’d kind of like to try the big company experience, and infrastructure and such. But yeah, you know, I…
15 00:07:32.520 ⇒ 00:07:33.899 Keith Saver: I’d been, …
16 00:07:34.530 ⇒ 00:07:42.430 Keith Saver: like, at the end of the day, I think for me, it was like, I had a couple other things that I wanted to work on, and Shopify doesn’t let you do anything on the side. …
17 00:07:42.730 ⇒ 00:07:48.139 Keith Saver: And so, I decided to take the leap. So, I’m actually trying to start two things right now.
18 00:07:48.140 ⇒ 00:07:48.570 Uttam Kumaran: Sure.
19 00:07:48.570 ⇒ 00:08:03.420 Keith Saver: both kind of part-time, so, you know. One of them, which I think I mentioned over text very briefly, is I’m starting a non-profit right now, and so the nonprofit is going to be focused on basically researching and emphasizing positive mental health benefits of video gaming.
20 00:08:03.540 ⇒ 00:08:19.640 Keith Saver: Okay. Which is obviously a little bit out of the standard discourse around video gaming. Yeah, so I, like, grew up as a pretty serious chess player, and I’ve played video games pretty competitively for my whole life. Okay. So it kind of comes from that hobby, and actually, I worked in a lot of, like, mental health care work, you know, data science-wise, too.
21 00:08:19.640 ⇒ 00:08:24.570 Keith Saver: So my, my girlfriend’s actually, starting to be at BCBA. She works with, … Oh!
22 00:08:24.570 ⇒ 00:08:39.120 Uttam Kumaran: Yeah, she works in ABA therapy, she… and… That’s awesome! I’ve learned, I… you know, my background is in computer engineering. I’ve… I’ve worked as a data engineer and done a bunch of data stuff, but traditional kind of engineering background. I’ve read, actually, you know, I’ve read
23 00:08:39.120 ⇒ 00:08:54.960 Uttam Kumaran: Fair bit of behavioral psychology, but mainly for fun. But when I met her, it kind of changed a lot about how I look at people, because there are a lot you can learn about people from learning about kids with, you know, autism or speech issues. Like, it tends to be…
24 00:08:54.960 ⇒ 00:09:12.020 Uttam Kumaran: You know, I think they tend to have the most exaggerated versions, but a lot of, like, what she tells me about them and behaviors is something that, you know, we all have in different ways, and yeah, just, like, really, really interesting field, like, field of behavior, you know, so definitely, that’s awesome.
25 00:09:12.440 ⇒ 00:09:28.050 Keith Saver: Yeah, yeah, absolutely. Yeah, it’s very cool. Yeah, so yeah, I’m starting that with this… my co-founder’s this guy who I actually worked with in my last job, so he’s a psychiatry professor at Columbia, and so there we work together on sort of, like, statistical modeling of how patients respond to opioid use disorder treatments and things like that.
26 00:09:28.050 ⇒ 00:09:35.230 Keith Saver: But, like, he’s also always been interested in this, so we’re starting… so that’s, like, one of my things, and then the other thing I’m trying to start, very different, is, like, a….
27 00:09:35.230 ⇒ 00:09:40.639 Uttam Kumaran: personal finance coaching business specifically aimed at doctors. Interesting!
28 00:09:40.640 ⇒ 00:09:55.830 Keith Saver: Yeah, a little all over the place. But that’s, like, it’s a big personal interest of mine, it’s something I’ve… and honestly, like, my, my partner’s, she’s an OBGYN, and so, like, I’ve kind of been doing this for a number of our friends for a number of years, and some sort of….
29 00:09:55.830 ⇒ 00:09:58.600 Uttam Kumaran: So you’re, like, what’s your personal finance stat?
30 00:09:59.080 ⇒ 00:10:04.889 Keith Saver: I mean, I just manage everything myself, like… Okay. I just, like, track things in Google Sheets, honestly.
31 00:10:04.890 ⇒ 00:10:05.430 Uttam Kumaran: Okay.
32 00:10:05.430 ⇒ 00:10:23.760 Keith Saver: It’s very simple. Okay. But yeah, I, you know, I think for doctors, it’s mostly, like, a lot of it is honestly behavioral, too. That, like, doctors have such a specific career trajectory and income trajectory that comes along with it, that there’s, like, a very specific, like, personal finance dynamic
33 00:10:23.930 ⇒ 00:10:26.959 Keith Saver: and behavioral dynamic that’s sort of intertwined there.
34 00:10:26.960 ⇒ 00:10:28.040 Uttam Kumaran: Oh, interesting.
35 00:10:28.040 ⇒ 00:10:33.139 Keith Saver: It’s like, yeah, yeah, it’s basically aimed to, like, the real short version of it is, like.
36 00:10:33.930 ⇒ 00:10:50.229 Keith Saver: you know, doctors are high performers their whole lives who are probably in 500,000 of debt when they graduate residency, and suddenly get a high income. And so they’ve had this delayed gratification, and they see all these people around them between other high-performing friends in other fields and slightly older doctors who are spending all this money.
37 00:10:50.230 ⇒ 00:10:59.659 Keith Saver: And there’s this, like, high pressure between both of those to, like, start spending some money, and or they do the responsible, like, the quote-unquote responsible thing and hire a financial advisor.
38 00:10:59.660 ⇒ 00:11:09.370 Keith Saver: And they don’t do the math on both, like, how much their first 5 years will impact their retirement from a compounding perspective, as well as, like.
39 00:11:09.420 ⇒ 00:11:10.220 Keith Saver: what?
40 00:11:10.570 ⇒ 00:11:17.280 Keith Saver: do financial advisor fees actually cost them? So I think there’s this, like, very important, like, 3-5 year window at the start of doctors, like.
41 00:11:17.920 ⇒ 00:11:26.939 Keith Saver: serious earning careers, where, like, a little bit of behavioral help and understanding of personal finance dynamics, like, is actually a multi-million dollar difference by the time they retire.
42 00:11:27.290 ⇒ 00:11:38.690 Uttam Kumaran: Yeah, I mean, there’s just, like, microcosms of this, you know, law, banking, like, you know, I’m sure you could easily tweak to serve multiple audiences, but that’s awesome. It’s also a data problem, you know?
43 00:11:38.690 ⇒ 00:11:40.080 Keith Saver: It is. Slightly.
44 00:11:40.670 ⇒ 00:11:49.550 Keith Saver: Yeah, so anyways, those are the two things I’m trying to start, but obviously those are, like, you know, both sort of part-time, so, you know, I was talking with Jamie and being like, yeah, I might…
45 00:11:49.730 ⇒ 00:11:55.060 Keith Saver: think about doing some part-time data science work, and that’s when he also mentioned, you and Brainforge.
46 00:11:55.450 ⇒ 00:12:08.659 Uttam Kumaran: Yeah, so, you know, Jamie’s a long-time friend of mine. I always… every time I talk to someone, I’m like, oh, I wish… I wish you were closer by, because we… we both worked at WeWork, and that was our first jobs together, so he… actually, he entered there.
47 00:12:08.890 ⇒ 00:12:09.429 Keith Saver: And we were….
48 00:12:09.430 ⇒ 00:12:23.230 Uttam Kumaran: We both graduated in 2018, and we were just thrown into, like, a… I think it was, like, a six- or seven-person data team that was running the data for the whole company. And so that team scaled from, like, six to, like.
49 00:12:23.230 ⇒ 00:12:25.830 Keith Saver: Wow, so you guys are really at the start, huh? Yeah, yeah, yeah.
50 00:12:25.830 ⇒ 00:12:37.289 Uttam Kumaran: Yeah, and, like, we were kind of… we joined when the company was 6,000 people, but most of WeWork is, like, 4,500 or 5,000 people were retail, so, like, it was about 1,000 corporate employees.
51 00:12:37.290 ⇒ 00:12:47.330 Uttam Kumaran: And it went into, like, about 15,000 people, when we sort of, like, when I left, and he sort of left shortly after, after 2020. Yeah, yeah. …
52 00:12:47.330 ⇒ 00:12:58.680 Uttam Kumaran: And yeah, I mean, we got to do, like… I still am basically not doing stuff more complicated than that, like… Oh, yeah. Complex in terms of, like, the amount that was at stake.
53 00:12:58.810 ⇒ 00:13:10.830 Uttam Kumaran: Like, the types of people I was dealing with, and, like, we were, we were, like, we basically had built our own sort of, like, dbt Cloud before dbt Cloud was there.
54 00:13:10.980 ⇒ 00:13:11.740 Keith Saver: We’re just….
55 00:13:11.740 ⇒ 00:13:14.149 Uttam Kumaran: Doing a lot of stuff in my 2019…
56 00:13:14.340 ⇒ 00:13:27.970 Uttam Kumaran: like, how do you run, like, a true, decentralized, like, data org? And it was sort of, like, just being written about then, which is really cool, and I was, like, 22 at the time, so it was, like, it was insane.
57 00:13:28.000 ⇒ 00:13:40.499 Uttam Kumaran: But, yeah, and then after that, I worked as a… the first data… data hire at this company called Flowcode. They’re based out of New York. They were, … they were a QR code startup, so I joined, actually, right the month before COVID.
58 00:13:40.640 ⇒ 00:13:41.230 Keith Saver: Oh, man.
59 00:13:41.230 ⇒ 00:13:50.840 Uttam Kumaran: And QR codes, like, really went crazy, and that company really blew up. It’s, like, very… got very lucky, but also, like, built the entire data stack.
60 00:13:51.050 ⇒ 00:14:06.849 Uttam Kumaran: by myself, and then got headcount, hired the data team, and then I also, after, like, kind of building all that up, I switched to go build our customer-facing data products, so I got into more product development, work as a product manager. And then I went to go….
61 00:14:09.600 ⇒ 00:14:10.590 Keith Saver: Whoop, hello?
62 00:14:13.670 ⇒ 00:14:16.010 Uttam Kumaran: Whoops.
63 00:14:16.010 ⇒ 00:14:19.729 Keith Saver: Sorry, I think cut off for a second, you said you, you were leaving Flowcode to go lead….
64 00:14:20.660 ⇒ 00:14:27.039 Uttam Kumaran: Oh, I was leading product at a data startup, also in York, that were two people that I met at WeWork, and then….
65 00:14:27.040 ⇒ 00:14:27.600 Keith Saver: I don’t know.
66 00:14:27.600 ⇒ 00:14:43.570 Uttam Kumaran: I moved here to Austin at that point. So that… I moved to Austin about 3 years ago, and then, yeah, started this business about 2 years ago. So we’re a data analytics and AI consultancy, primarily doing data modeling and data engineering work now, and doing…
67 00:14:43.800 ⇒ 00:14:46.670 Uttam Kumaran: Building AI agents for several clients.
68 00:14:48.130 ⇒ 00:14:50.200 Keith Saver: Very cool. Yeah, definitely makes sense.
69 00:14:50.530 ⇒ 00:14:54.190 Keith Saver: That’s awesome. Yeah, it’s a… it’s very cool, yeah, it’s starting from the….
70 00:14:54.190 ⇒ 00:14:56.370 Uttam Kumaran: I was sort of talking to Jamie a little bit at…
71 00:14:57.900 ⇒ 00:15:00.139 Uttam Kumaran: Yeah, I was… I was talking to Jamie…
72 00:15:00.150 ⇒ 00:15:19.089 Uttam Kumaran: just kind of… I always ask a certain couple of friends of mine, like, hey, who’s… who do you know that’s, could be interested in doing part-time or just data work in a consulting manner? And yeah, he gave me your contact, so that’s kind of, like, how… roundabout story, about how we kind of got connected, so….
73 00:15:21.110 ⇒ 00:15:25.200 Keith Saver: Yeah, that’s awesome, that’s very cool. So yeah, you’re up to, like, 15 people now or so?
74 00:15:25.200 ⇒ 00:15:28.289 Uttam Kumaran: Yes, we have about 15 people, which is really quite great.
75 00:15:28.290 ⇒ 00:15:29.360 Keith Saver: Very exciting.
76 00:15:29.360 ⇒ 00:15:32.479 Uttam Kumaran: Yeah, it’s been really, really cool to see.
77 00:15:32.620 ⇒ 00:15:41.890 Uttam Kumaran: you know, see and manage, like, all the growth. And it’s completely bootstrapped, the company, so I haven’t raised any money for it, just been slowly, so I’m trying to get clients and growing, so….
78 00:15:42.410 ⇒ 00:15:52.699 Keith Saver: I feel like, … I mean, it’s interesting, you know, I’ve always been thinking about this too, it’s like, I really prefer the professional services business model, I feel like, compared to the VC.
79 00:15:53.140 ⇒ 00:15:57.489 Keith Saver: Like, the professional services lifestyle, I feel like, also has some benefits.
80 00:15:58.120 ⇒ 00:16:01.620 Uttam Kumaran: I think it depends on how ambitious you are.
81 00:16:03.520 ⇒ 00:16:08.869 Uttam Kumaran: Like, my life.
82 00:16:09.160 ⇒ 00:16:20.340 Uttam Kumaran: is really in… in a tough spot right now, just because we want to grow.
83 00:16:20.490 ⇒ 00:16:24.769 Uttam Kumaran: Like, had I sort of slowly had…
84 00:16:25.400 ⇒ 00:16:32.260 Uttam Kumaran: You know, we already worked… we have about… 10 or 11…
85 00:16:34.720 ⇒ 00:16:41.979 Uttam Kumaran: Whoop! Hello. We’re… we’re continuing to grow pretty fast.
86 00:16:42.120 ⇒ 00:16:45.200 Uttam Kumaran: … Can you hear me? Am I cutting out again?
87 00:16:45.200 ⇒ 00:16:47.190 Keith Saver: It caught up for a second, I think you’re back now.
88 00:16:56.360 ⇒ 00:16:57.600 Keith Saver: Ho-hello?
89 00:16:59.820 ⇒ 00:17:01.080 Uttam Kumaran: Hey, can you hear me now?
90 00:17:01.080 ⇒ 00:17:01.620 Keith Saver: Yeah.
91 00:17:02.260 ⇒ 00:17:03.080 Uttam Kumaran: Okay.
92 00:17:03.400 ⇒ 00:17:06.170 Uttam Kumaran: Sorry, I’m just sitting down.
93 00:17:06.170 ⇒ 00:17:06.849 Keith Saver: No worries.
94 00:17:06.859 ⇒ 00:17:15.679 Uttam Kumaran: I think the professional services business model is good in that you can get cash in and cash out, but it’s tough to scale.
95 00:17:15.679 ⇒ 00:17:17.210 Keith Saver: Yes, definitely.
96 00:17:17.210 ⇒ 00:17:36.109 Uttam Kumaran: And it’s not… and the multiples are much different than, you know, a SaaS business, so I sort of have to contend with that a little bit, but I don’t know, I also felt like there’s a lot of sort of bad consultants. Like, I think the market of talent… the market of companies is not that great. Like, there’s not a lot of great consultancies.
97 00:17:36.730 ⇒ 00:17:43.910 Uttam Kumaran: Additionally, there’s only… there’s a lot of people that start, like, one or two-person consultancies, but they never get to, like… they never…
98 00:17:44.140 ⇒ 00:17:56.309 Uttam Kumaran: like, cross the… the gap of, like, actually scaling into something that is bigger than just… just them. Like, now, when I pitch us, I pitch Brainforge. Like, you can’t really, like… I’m… I am a…
99 00:17:56.460 ⇒ 00:18:01.130 Uttam Kumaran: I’m on the team, but, you know, we’re pitching the company, and that was a huge sort of thing to…
100 00:18:01.450 ⇒ 00:18:05.770 Uttam Kumaran: leap to make, you know, but it’s… everything’s hard in business, I don’t think, like…
101 00:18:06.650 ⇒ 00:18:08.880 Uttam Kumaran: I don’t know, maybe that’s just me, but, like, we set our.
102 00:18:08.880 ⇒ 00:18:09.660 Keith Saver: I think so.
103 00:18:09.660 ⇒ 00:18:11.049 Uttam Kumaran: Super high, and, like.
104 00:18:11.540 ⇒ 00:18:21.110 Uttam Kumaran: we’re trying to be the best there is in what we do, and the AI stuff that we do now has just pulled a lot of that forward, too, so… it’s… it’s been… it’s been a cool ride.
105 00:18:21.910 ⇒ 00:18:24.239 Keith Saver: Definitely. Yeah, that’s… that’s awesome.
106 00:18:25.810 ⇒ 00:18:32.369 Keith Saver: I feel like, yeah, scaling to that many people is… is a lot of work as a consultancy, I can only imagine.
107 00:18:32.370 ⇒ 00:18:41.399 Uttam Kumaran: Yeah, and it’s also like, you know, it’s also everything beyond the execution, right? It’s sales, it’s marketing, it’s operations, there’s finance.
108 00:18:41.530 ⇒ 00:18:45.530 Keith Saver: There’s partnerships now. Yeah, it’s just an absolute nightmare.
109 00:18:45.530 ⇒ 00:18:46.570 Uttam Kumaran: …
110 00:18:46.950 ⇒ 00:19:00.470 Uttam Kumaran: But, you know, I don’t have a boss, I haven’t had a boss in a couple years now. I can run, like, a… our team is really engineering heavy, so we try to make it a great place for engineers. That wasn’t the case in a lot of
111 00:19:00.710 ⇒ 00:19:03.989 Uttam Kumaran: the companies I was at. …
112 00:19:04.210 ⇒ 00:19:12.690 Uttam Kumaran: And yeah, just try to do this, like, honestly, like, try to actually deliver great data work, and work with awesome people, and, I think, like, we will start to hit, sort of, like.
113 00:19:12.750 ⇒ 00:19:21.390 Uttam Kumaran: some of the economies of scale here soon, where, you know, our OPEX isn’t rising super linearly toward, like, with our
114 00:19:21.400 ⇒ 00:19:33.179 Uttam Kumaran: you know, revenue that’s coming in. And also, like, our work is actually getting easier to do because of AI, so everybody on my team has Cursor, like, we tend to go to companies that have a lot of the same playbooks that we run, so…
115 00:19:33.250 ⇒ 00:19:52.790 Uttam Kumaran: like, our delivery costs should be going down as well. I think continuously, and, you know, kind of the reason why we’re on the call is just finding great people tends to be really, really tough. And, you know, we found an edge, I think mainly just because I worked in data for a while, and I know where to source people from, like, other angles, but, …
116 00:19:52.940 ⇒ 00:19:58.699 Uttam Kumaran: yeah, I think that’s… the people are both the best part and the most challenging part, you know?
117 00:19:58.700 ⇒ 00:19:59.690 Keith Saver: Definitely.
118 00:19:59.760 ⇒ 00:20:19.470 Uttam Kumaran: like, managing teams and building up, like, you know, we’re kind of like a decentralized data team, but we’re managing, like, 10 different clients that they don’t all… they’re not all in the… they’re not the same company, you know? So, just a… just, like, on… just a different level of, of complexity, I feel like. But it’s been fun, yeah, it’s been good.
119 00:20:20.820 ⇒ 00:20:22.150 Keith Saver: Yeah, absolutely.
120 00:20:22.610 ⇒ 00:20:36.540 Uttam Kumaran: Yeah, so I’m kind of curious on, like, you know, what you’re thinking about short-term. I mean, we’re sort of bringing on a couple more data folks. I’m kind of interested to hear a little bit about, like, what you’d be interested in doing, or…
121 00:20:36.570 ⇒ 00:20:49.109 Uttam Kumaran: If you kind of, like, you know, have any interest in joining and doing some part-time work for a company like ours. And yeah, just, like, wanted to just have a kind of a cash conversation, so happy to…
122 00:20:49.170 ⇒ 00:20:51.540 Uttam Kumaran: You know, answer any questions on my end, too.
123 00:20:51.920 ⇒ 00:21:00.509 Keith Saver: Yeah, no, absolutely. I think, … yeah, in general, I’d say, like, you know, I haven’t locked it down exactly. I’d say I’d, in general, probably be…
124 00:21:00.730 ⇒ 00:21:04.910 Keith Saver: Looking for, like, in the ballpark of 1-2 days of work a week.
125 00:21:04.910 ⇒ 00:21:05.280 Uttam Kumaran: Okay.
126 00:21:05.280 ⇒ 00:21:08.820 Keith Saver: To maintain enough time, for my other projects.
127 00:21:09.000 ⇒ 00:21:27.380 Keith Saver: But, like, I would definitely be interested. You know, I’ve… obviously, like, those have some data work, but, like, I also like data work and want to keep my skills sharp, and, like, I, you know, I’ve done some, you know, I’ve done more, sort of, consulting-style client work at some roles in the past, and I like that too, and I think that, like.
128 00:21:28.650 ⇒ 00:21:44.520 Keith Saver: you know, to be honest, I think, like, one of my minor fears when leaving Shopify was I was like, I want to make sure that I’m getting enough AI exposure still. Yeah. And, like, and so that’s, you know, one place both in my other projects and in consulting that I’m definitely interested in, so hearing your emphasis there is certainly interesting.
129 00:21:44.520 ⇒ 00:21:54.399 Keith Saver: I’d be curious also to hear, like, what types of work. I think, like, to be realistic with myself, like, I think I both am better at and enjoy more of the, like.
130 00:21:54.400 ⇒ 00:22:04.339 Keith Saver: product analytics and strategy, like, product strategy… strategy work more than some of the data engineering work. Like, I can do some of it, but it’s not my favorite.
131 00:22:04.810 ⇒ 00:22:21.840 Uttam Kumaran: Yeah, so… so we kind of are… do, like, full-stack data, so we come in and we do data engineering, we do data modeling, we also do reporting and BI, and then we also have a real strong, you know, presence, in product analytics, so we do a lot of work with amplitude, mixed panel, segments.
132 00:22:21.920 ⇒ 00:22:34.850 Uttam Kumaran: Router stack, and, like, mainly a lot of our work is trying to drive towards outcomes for some of these businesses, so whether they’re missing data about users, they launch new business lines, they can’t report on it.
133 00:22:34.850 ⇒ 00:22:44.909 Uttam Kumaran: So we tend to come in and sort of isolate, like, what is the challenge? The clients sometimes don’t know, and these are, like, mid-market businesses, so typically, like, 20 to 100 million in revenue.
134 00:22:44.910 ⇒ 00:22:56.399 Uttam Kumaran: So we’re not working with, like, startups, but we’re also, like, slowly kind of growing into the size of companies. So really, like, we come in and we can make a lot of procurement decisions, bring in the right tools, and then sort of deploy
135 00:22:56.400 ⇒ 00:23:07.069 Uttam Kumaran: you know, outcomes to help them make decisions, whether it’s, like, reverse ETL of models back into systems, whether it is a dashboard, you know, but we are doing a lot of
136 00:23:07.070 ⇒ 00:23:16.979 Uttam Kumaran: product analytics work, so a lot of, like, event tagging, understanding funnels, you know, sharing, sort of, like, what changes should be made. So we are doing quite a bit of that type of work.
137 00:23:16.980 ⇒ 00:23:21.910 Keith Saver: Okay, yeah. No, I mean, especially in that type of work, I think I’d be very interested. ….
138 00:23:21.910 ⇒ 00:23:38.340 Uttam Kumaran: Yeah, I was… maybe you could talk a little bit about, like, some of the work you’ve done in that area, or anywhere, like, sort of what type of work you were doing at Shopify, or even before, and just so I can get a sense of, like, the types of, analytics that you were working on.
139 00:23:39.290 ⇒ 00:23:47.490 Keith Saver: Yeah, absolutely. Maybe I’ll actually give… I can give some examples from earlier work. You said you have a… I was looking through your website, and I saw your text, too. You have a chunk of, like, healthcare clients, too, right?
140 00:23:47.490 ⇒ 00:23:48.090 Uttam Kumaran: Yes.
141 00:23:48.090 ⇒ 00:24:08.369 Keith Saver: talk to some of that. Yeah, because, like, that’s honestly a big interest of mine. I have some experience there. So, before I was at Shopify, I worked at this company, Paratherapeutics. So, Paratherapeutics, was this company making prescription digital therapeutics, for treating mental health conditions. So, it was basically a couple products. One was for insomnia, and one was for,
142 00:24:08.410 ⇒ 00:24:20.720 Keith Saver: substance use disorder, in particular focus on opioid use disorder. And so Pear was basically, like, actually the first company to basically get these products approved through these new softwares and medical device pathways through the FDA.
143 00:24:20.940 ⇒ 00:24:31.190 Keith Saver: And so there, I was, like, one of the, you know, first couple data scientists, and so I sort of, you know, led product analytics. I particularly for the Insomnia product. I…
144 00:24:31.420 ⇒ 00:24:36.920 Keith Saver: was sort of, like, the lead data scientist on it as we went commercial. So, you know, it was kind of a mix of, like.
145 00:24:37.310 ⇒ 00:24:45.529 Keith Saver: You know, partnering with engineering, designing instrumentation, you know, partnering with data engineering, designing all the data models, building all the dashboards, …
146 00:24:45.530 ⇒ 00:25:10.230 Keith Saver: you know, doing A-B testing on product improvements and that sort of thing. One of the other interesting projects I worked on at PAIR, this is actually on the other product, on the addiction product, was there was basically, it’s called contingency management. It’s basically a monetary reward system for getting patients to engage with therapy. So in that context, I also did a lot of work around analyzing, like, patient engagement and behavioral patterns in the context of these
147 00:25:10.230 ⇒ 00:25:21.639 Keith Saver: long-term rewards, which were given in a probabilistic manner, and sort of optimizing that algorithm within, obviously, a variety of company constraints across, you know, compliance, legal, engineering, etc. And that’s a very high level, but to give you a quick…
148 00:25:21.770 ⇒ 00:25:22.740 Keith Saver: Quick summary.
149 00:25:23.310 ⇒ 00:25:34.299 Uttam Kumaran: Yeah, so we, you know, we actually just onboarded another, I think, telehealth company today. They’re just starting today. We also work with some sort of prescription medication companies.
150 00:25:34.300 ⇒ 00:25:45.539 Uttam Kumaran: That are doing, like, GLP-1 or other medications, sort of, like, online now that, like, I think some of the laws have changed. And then, yeah, we’ve worked with a couple of different, like, healthcare-related companies, so that’s awesome.
151 00:25:45.540 ⇒ 00:25:57.300 Uttam Kumaran: I mean, in terms of, like, on the product analytics side, like, if you could talk nitty-gritty, I mean, we have some work that’s, like, as narrow as, like, hey, we just need to establish new product events, like, what are the events we want to track?
152 00:25:57.350 ⇒ 00:26:04.570 Uttam Kumaran: I’m also, of course, all the way high level to, like, making changes to, like, online products or online flows, like.
153 00:26:04.640 ⇒ 00:26:06.769 Uttam Kumaran: Where in there do you think you would…
154 00:26:06.900 ⇒ 00:26:10.219 Uttam Kumaran: Feel most interested, but also, like, most capable, you think?
155 00:26:10.770 ⇒ 00:26:18.859 Keith Saver: I can do a mix. I’d say, in general, I sort of enjoy some of the higher-level flow more, but, like, I… obviously, you have to instrument, like… Yeah.
156 00:26:18.860 ⇒ 00:26:33.480 Keith Saver: You know, that’s… and it’s, like, I’m more than happy to do that. At Shopify, I actually did a lot more of the lower level designing instrumentation, of, like, you know, partnering with our engineering teams, like, as new product… products or product features are being built, and, like, designing our instrumentation processes from there.
157 00:26:33.830 ⇒ 00:26:40.019 Uttam Kumaran: Yeah, the reason I ask is, like, we probably would be looking for someone that could do a little bit of both, like, there is some instrumentation.
158 00:26:40.120 ⇒ 00:27:00.080 Uttam Kumaran: But it’s… these are not, like, Shopify-level products, like, usually it’s, like, you can get it all done in a couple sprints, and then… and then it quickly moves into, like, okay, like, can we measure and actually start to give out recommendations? Like, one of the challenges in our business is, like, it’s been hard for us to, like, sort of hire for people that can just do one thing, like.
159 00:27:00.760 ⇒ 00:27:14.049 Uttam Kumaran: We need some overlap in skill set, but also overlap in, like, you can kind of really wear multiple hats, but not in the sense of, like, oh, you need to now go do marketing stuff. It’s more of, like, within a given, like.
160 00:27:14.260 ⇒ 00:27:24.739 Uttam Kumaran: within a given, you know, discipline like product analyst, can you jump between reporting out, giving recommendations, working with an engine to instrument, making sure those fire, like.
161 00:27:24.760 ⇒ 00:27:43.000 Uttam Kumaran: it’s kind of like that, which, you know, it’s, again, more just hypotheses on how we’re trying to hire and recruit people longer term, is, like, getting people that can jump, you know, across both the low-level stuff and sort of work directly with clients. We found that people that are just isolated to just, like, do DBT models.
162 00:27:43.060 ⇒ 00:27:48.709 Uttam Kumaran: Yeah, we could do that, but actually, like, the best… we can’t have… we don’t have, like, 10 people on every client, so…
163 00:27:49.130 ⇒ 00:27:54.020 Uttam Kumaran: we need people that can kind of rotate a little bit, so I actually think there is probably an opportunity to get some of the help.
164 00:27:54.100 ⇒ 00:27:59.629 Uttam Kumaran: You know, that you’re talking about on a couple clients. I mean, in terms of, like, timing.
165 00:27:59.640 ⇒ 00:28:18.229 Uttam Kumaran: Yeah, I think the… probably my only, like, feedback would be, like, some of these clients are… they’re ongoing, meaning work happens every day, so I guess what we’d be more interested in is, like, are you able to dedicate, like, even a couple hours every day to some of this work, or do you think it would be more isolated to, like, I can only do, like, Mondays and Tuesdays, and then we’re sort of off?
166 00:28:19.180 ⇒ 00:28:21.179 Keith Saver: I could have flexibility with that.
167 00:28:21.390 ⇒ 00:28:24.189 Keith Saver: Okay. Like, I don’t have anything set in stone on that. I could definitely be fine.
168 00:28:24.190 ⇒ 00:28:28.559 Uttam Kumaran: Okay. Okay. So yeah, I mean, I think if you’re sort of like, hey, I’m open to, like.
169 00:28:28.770 ⇒ 00:28:48.610 Uttam Kumaran: like, 10 to 20 hours a week, then that’s something that probably we can look for. I think anything less than 10 hours can be tricky, just because I’m not… we’re not 100% sure on the scope yet, but, like, usually for anyone part-time, like, that’s sort of what we expect. And then also, this isn’t, like… usually everything in terms of meetings and stuff is pretty…
170 00:28:48.610 ⇒ 00:29:03.090 Uttam Kumaran: like, low. I think for most clients, we typically do, like, every other day, stand-ups, but also, it’s, again, it’s up to the project manager, so there may be grooming and stuff like that, but we’re completely a… we’re completely remote, like, mostly async companies, so…
171 00:29:03.090 ⇒ 00:29:09.999 Uttam Kumaran: But it’s a lot of writing, it’s a lot of Slack, it’s a lot of Notion, so it’s really the meetings only increase of, like.
172 00:29:10.310 ⇒ 00:29:21.260 Uttam Kumaran: communication is happening. In fact, like, I just… we do so much writing that I think that tends to take care of a lot of stuff. And then it’s meeting with clients that tend to be the other sort of lock of time. …
173 00:29:21.820 ⇒ 00:29:37.249 Uttam Kumaran: But yeah, maybe, maybe one thing I can do is I can actually also introduce you to one other person on the team, my business partner, Robert. He’s really heavy on the product analytics side. I would love for you to maybe meet him, and just… I’ll tell him to just give you a little bit of an overview of a couple of the clients that we have.
174 00:29:37.330 ⇒ 00:29:51.090 Uttam Kumaran: I think someone with your background, too, like, I would love to have you more at the higher level. We have some people that can do some of the instrumentation work, but I think we’re lacking some of the folks that can actually translate, like, sort of be a solution architect on the
175 00:29:51.250 ⇒ 00:30:07.249 Uttam Kumaran: product analytics side, right? So, like, help establish roadmaps, work with PMs to create those tickets, and then actually, like, you know, deliver… actually explain to clients, like, what it is we’re even doing. Like, that’s where I think you could fit in really, really nicely in sort of a part-time manner on maybe one or two clients.
176 00:30:07.630 ⇒ 00:30:12.059 Keith Saver: Yeah, I would definitely be… I’d be very interested in that kind of work. That’s a lot of the work I enjoy the most. So, definitely.
177 00:30:12.060 ⇒ 00:30:22.010 Uttam Kumaran: Okay, cool. And then, yeah, on the AI side, look, we’re doing a ton of stuff, like, we’re trying to find ways to marry AI with a lot of the work that we’re doing on the data side. Like, for example.
178 00:30:22.010 ⇒ 00:30:33.270 Uttam Kumaran: as part of a lot of our work, we tend to be… we tend to do a lot of writing, we record a lot of meetings, we have access to Slack, so we actually tend to have logs, you know, of decisions that get made, and
179 00:30:33.330 ⇒ 00:30:41.650 Uttam Kumaran: And I think there’s gonna be an opportunity for us to use AI, plus a lot of the data we’re getting from systems like product analytics, but also our entire warehouse, to actually help
180 00:30:41.740 ⇒ 00:30:58.720 Uttam Kumaran: our business users make decisions, or even help our team, you know, use agents to ask questions about data and… versus, like, just start writing SQL and start to do discovery. So, there is some innovation that is to do there. I mean, again, we’re not a product company, so…
181 00:30:58.750 ⇒ 00:31:04.940 Uttam Kumaran: Where innovation kind of happens, like, with whatever time remaining, but… We’re trying to push, because…
182 00:31:05.020 ⇒ 00:31:21.399 Uttam Kumaran: we’re coming to a lot of these companies that just… they don’t… they’re never gonna establish, like, a full data team with, like, hundreds of people that can go dissect every nuance, but I think there’s ability for AI to do that now. And so, on the AI side, we… we do a lot of agent building, we build agent workflows for several clients, …
183 00:31:21.400 ⇒ 00:31:31.240 Uttam Kumaran: And so there’s some route where I’m trying to marry that team and have them learn from what the data team is doing and build, sort of, workflows. So even as you’re working on stuff.
184 00:31:31.280 ⇒ 00:31:42.009 Uttam Kumaran: like, if we’re able to create SOPs and things around common workflows, I think there’s certainly ways for us to find out how AI can… can augment some of that. So, I think it’d be interesting, I think you’d scratch that itch as well.
185 00:31:42.500 ⇒ 00:31:48.229 Keith Saver: Yeah, yeah, that’d be awesome. Yeah, it sounds great, I’d love to chat with your business partner. If it’s helpful, I can also send you a resume.
186 00:31:48.800 ⇒ 00:31:55.430 Uttam Kumaran: Yeah, resume would be great, and then I’ll probably coordinate over email, and then get you, like, hooked up with him probably sometime early next week.
187 00:31:55.880 ⇒ 00:31:57.979 Keith Saver: Okay. Yeah, absolutely. That sounds great.
188 00:31:57.980 ⇒ 00:32:00.359 Uttam Kumaran: Okay, dope. Any other questions for me?
189 00:32:00.800 ⇒ 00:32:07.570 Keith Saver: … I don’t think so right now. I’ll shoot you an email if I come up with some. But I think, I think that makes sense. That sounds great.
190 00:32:07.570 ⇒ 00:32:13.509 Uttam Kumaran: Yeah, and if I can be helpful for anything on your nonprofit stuff, or any of the other ideas, let me know. I’m happy to help.
191 00:32:13.740 ⇒ 00:32:21.780 Keith Saver: Yeah, no, I mean, I’d love to, some other point, probably pick your brain as I’m trying to wade through all the, like, you know, starting a business side.
192 00:32:23.100 ⇒ 00:32:24.360 Uttam Kumaran: It’s an unfortunate size.
193 00:32:24.360 ⇒ 00:32:26.040 Keith Saver: Absolutely. Yeah.
194 00:32:27.850 ⇒ 00:32:30.660 Uttam Kumaran: Okay, awesome. Well, thank you, Keith, I appreciate it.
195 00:32:30.660 ⇒ 00:32:32.520 Keith Saver: Yeah, absolutely, thank you too, really appreciate it.
196 00:32:32.520 ⇒ 00:32:34.159 Uttam Kumaran: Okay, great, talk to you soon.
197 00:32:34.160 ⇒ 00:32:34.770 Keith Saver: Yeah.