Meeting Title: Demo Dry-run Date: 2026-01-08 Meeting participants: Pranav, Samuel Roberts, Uttam Kumaran
WEBVTT
1 00:00:54.520 ⇒ 00:00:55.620 Samuel Roberts: Nay.
2 00:00:56.330 ⇒ 00:00:57.950 Pranav: Yo, yo, how are you feeling?
3 00:01:00.540 ⇒ 00:01:01.230 Pranav: Oh, man.
4 00:01:01.230 ⇒ 00:01:04.950 Samuel Roberts: Not… not better, I’ll put it that way, at least. It’s not…
5 00:01:05.110 ⇒ 00:01:06.810 Samuel Roberts: I don’t know. I don’t know. I just…
6 00:01:07.270 ⇒ 00:01:13.089 Samuel Roberts: Cold is the main symptom right now, but I don’t know if I’m just getting tired, or… worse, you know?
7 00:01:13.660 ⇒ 00:01:15.550 Pranav: Fair. Yeah, yeah. I mean…
8 00:01:15.800 ⇒ 00:01:25.290 Pranav: I think you should take it a little easy now. That, that demo was really good. Yeah, I’m just trying to get excited about that, too. But yeah, no, I’m happy with that. I was…
9 00:01:25.290 ⇒ 00:01:27.530 Samuel Roberts: the reason I didn’t want to deploy it
10 00:01:28.000 ⇒ 00:01:35.649 Samuel Roberts: because I was trying to figure out, since we have a front-end, a back-end, a database, and four MCP servers.
11 00:01:35.650 ⇒ 00:01:36.310 Pranav: Yeah.
12 00:01:37.040 ⇒ 00:01:42.450 Samuel Roberts: I want to make sure that not only are those URLs gonna go kind of automatically.
13 00:01:42.580 ⇒ 00:01:45.269 Samuel Roberts: Into the… where they need to be.
14 00:01:46.080 ⇒ 00:01:48.840 Samuel Roberts: But also, all of the…
15 00:01:50.370 ⇒ 00:01:54.670 Samuel Roberts: the keys and stuff for Google and Meta, because they’re apps.
16 00:01:55.660 ⇒ 00:01:56.530 Pranav: Right.
17 00:01:56.860 ⇒ 00:02:03.439 Samuel Roberts: Not just the… yeah, so, like, I just… I’m trying to figure that out, and having… having cursor, I can just, like…
18 00:02:03.780 ⇒ 00:02:05.670 Samuel Roberts: Push all those keys up.
19 00:02:08.850 ⇒ 00:02:14.770 Pranav: Yeah. I also don’t think you need to worry too much about, like, a little bit of downtime, either. If it’s easier to kind of just…
20 00:02:15.130 ⇒ 00:02:16.230 Pranav: Commit them.
21 00:02:16.330 ⇒ 00:02:17.760 Pranav: And then.
22 00:02:18.710 ⇒ 00:02:25.770 Samuel Roberts: Yeah, I guess that’s a good point. I was just trying to think, like, as we start pushing this and we need to set up other environments, I want to kind of not have to be…
23 00:02:25.900 ⇒ 00:02:28.019 Samuel Roberts: Digging through the dashboard the whole time.
24 00:02:28.550 ⇒ 00:02:33.350 Samuel Roberts: Yeah, so right now, we’re working in a kind of weird way, because…
25 00:02:33.350 ⇒ 00:02:37.919 Pranav: they are using the development URL and not the production URL.
26 00:02:38.200 ⇒ 00:02:38.570 Samuel Roberts: Yeah.
27 00:02:38.570 ⇒ 00:02:40.469 Pranav: And so…
28 00:02:40.640 ⇒ 00:02:46.849 Pranav: I honestly don’t see there being much of an issue, like, if the development, like, instance goes down.
29 00:02:47.010 ⇒ 00:02:51.399 Pranav: Because they’ll be looking at production. Production is just the one that shouldn’t be going down.
30 00:02:51.400 ⇒ 00:02:59.190 Samuel Roberts: No, I agree, I agree. I’m not too worried about that, I just… I want to make sure, since we’re adding all these MC… we’re adding 3 more services, that I get the URLs right, and that it…
31 00:02:59.190 ⇒ 00:02:59.780 Pranav: Clear.
32 00:02:59.810 ⇒ 00:03:02.770 Samuel Roberts: Because then moving to prod, it’s going to be the same thing.
33 00:03:03.920 ⇒ 00:03:08.100 Samuel Roberts: I just wanted to make sure that there was a clean way to do this, and I think I got it.
34 00:03:08.100 ⇒ 00:03:08.900 Pranav: Oh, boy.
35 00:03:09.120 ⇒ 00:03:12.870 Samuel Roberts: I’m just testing on a… I made a new environment called Sam Test.
36 00:03:13.010 ⇒ 00:03:18.169 Samuel Roberts: And I’m trying to sync to that to make sure I don’t, like, just overwrite everything, but I think we’ll be pretty good.
37 00:03:18.630 ⇒ 00:03:19.490 Pranav: Cool, cool.
38 00:03:21.290 ⇒ 00:03:27.150 Pranav: So, in that… yeah, I just saw that meeting randomly come up in the external chat. Was that, just kind of…
39 00:03:28.460 ⇒ 00:03:31.810 Pranav: like, did Bobby ask for that, or was Utsu.
40 00:03:31.810 ⇒ 00:03:35.500 Samuel Roberts: Yeah, I’m trying to remember now, sorry, my brain’s a little…
41 00:03:35.650 ⇒ 00:03:36.590 Pranav: Boggy.
42 00:03:36.730 ⇒ 00:03:42.130 Samuel Roberts: Yeah, they had asked for something, just trying to… I think they were… Wanting to basically, like.
43 00:03:44.700 ⇒ 00:03:54.460 Samuel Roberts: you know, speed up, obviously, but, like, optimize what we’re working on. So the result was basically that they want to be able to have this forecasting tool
44 00:03:54.780 ⇒ 00:03:59.859 Samuel Roberts: And not necessarily have all the historical data, but just be able to, like, input stuff and have it forecast.
45 00:04:00.510 ⇒ 00:04:01.230 Pranav: Gotcha.
46 00:04:01.230 ⇒ 00:04:01.970 Samuel Roberts: So.
47 00:04:01.970 ⇒ 00:04:04.759 Pranav: So it’s more just like a calculator, just like a front-end tool.
48 00:04:04.970 ⇒ 00:04:08.829 Samuel Roberts: For now, yeah. So, but obviously, like, as, you know…
49 00:04:09.360 ⇒ 00:04:12.050 Samuel Roberts: you and I need to make sure that we’re ready for, you know.
50 00:04:12.240 ⇒ 00:04:15.859 Samuel Roberts: being able to input the data, not just by the UI, you know what I mean?
51 00:04:16.310 ⇒ 00:04:17.200 Pranav: Gotcha, yep.
52 00:04:17.200 ⇒ 00:04:18.190 Samuel Roberts: So I think, like.
53 00:04:18.360 ⇒ 00:04:29.409 Samuel Roberts: It saves us a little bit, because we can just focus on that for now, and not have to worry about that. But I think next week, either you, or we might loop in Casey, may start on some of the data stuff.
54 00:04:30.400 ⇒ 00:04:32.339 Samuel Roberts: Depending on how it all works out.
55 00:04:33.680 ⇒ 00:04:38.839 Pranav: Cool, cool. Yeah, I was just actually talking to Casey about, the Slack reporting stuff.
56 00:04:39.020 ⇒ 00:04:39.530 Samuel Roberts: Oh, cool.
57 00:04:39.910 ⇒ 00:04:51.959 Pranav: Yeah, the Slack reports look very bare bones in terms of, like, what the previous agency had set up. Yeah. It’s straight up just, like, a table, and I think it’s just a… it’s an image of a table.
58 00:04:52.740 ⇒ 00:04:53.819 Samuel Roberts: Oh, really?
59 00:04:53.820 ⇒ 00:04:58.439 Pranav: Yeah, so I think we can do a little bit better there, have an actual table in Slack.
60 00:04:58.840 ⇒ 00:05:01.760 Samuel Roberts: Yeah. Have you done anything with the Slack API much?
61 00:05:02.770 ⇒ 00:05:09.310 Pranav: I think I might have, like, a while ago. Okay.
62 00:05:09.310 ⇒ 00:05:10.760 Samuel Roberts: I haven’t done a ton.
63 00:05:10.760 ⇒ 00:05:11.190 Pranav: Yeah.
64 00:05:11.190 ⇒ 00:05:17.610 Samuel Roberts: I know… yeah, Casey Mustafa did a little bit. I know what we looked at with one other client when we were
65 00:05:17.770 ⇒ 00:05:22.810 Samuel Roberts: like, doing, like, a Slack interface was, like, the text versus the…
66 00:05:23.110 ⇒ 00:05:25.030 Samuel Roberts: Is it block that they call them?
67 00:05:25.340 ⇒ 00:05:26.630 Pranav: Yeah, blocks.
68 00:05:26.630 ⇒ 00:05:33.159 Samuel Roberts: Yeah, okay. So I just don’t… yeah, I’m not sure, I don’t know that much, but that was all I was just gonna be like, yeah, blocks. That’s all I got for you right now.
69 00:05:33.160 ⇒ 00:05:37.500 Pranav: Okay, cool, yeah, I mean, yeah, Casey, like…
70 00:05:37.710 ⇒ 00:05:43.139 Pranav: he helped me out a lot. He kind of, like, told me, like, all the stuff that he’s done. He sent me, like…
71 00:05:43.370 ⇒ 00:05:49.460 Pranav: A… a bot that he made, probably for a different client or internally, so…
72 00:05:49.690 ⇒ 00:05:51.159 Pranav: I’ll be taking a look at that.
73 00:05:52.230 ⇒ 00:05:52.860 Samuel Roberts: Right.
74 00:05:53.270 ⇒ 00:06:01.149 Pranav: Yup so you kind of already did the demo today… With them, right?
75 00:06:02.820 ⇒ 00:06:06.669 Pranav: like, in that call, did you, like, also just, like, show them the MCP servers, or was it more so.
76 00:06:06.670 ⇒ 00:06:12.910 Samuel Roberts: No, no, no, I had actually shut everything down because I was… my computer was starting to, be very, very slow.
77 00:06:12.910 ⇒ 00:06:13.960 Pranav: Gotcha, okay.
78 00:06:13.960 ⇒ 00:06:19.260 Samuel Roberts: And cursor was taking forever, so I didn’t do that, but I… yeah, the lume was the only thing I got. Okay.
79 00:06:19.260 ⇒ 00:06:22.160 Pranav: Gotcha, gotcha. Okay, cool. Well,
80 00:06:23.240 ⇒ 00:06:33.989 Pranav: let’s see, for tomorrow’s demo, I feel pretty good. I’m gonna have, like, a V1 completed, probably for, like, the Slack reporting.
81 00:06:34.390 ⇒ 00:06:35.030 Samuel Roberts: Boom.
82 00:06:35.030 ⇒ 00:06:39.609 Pranav: Basically, what it’s gonna look like is just it’s gonna go to, like, one of our channels, and it’s just gonna be, like.
83 00:06:39.720 ⇒ 00:06:47.960 Pranav: basically, it’ll… I’ll have it working with, like, sample data to begin with,
84 00:06:48.450 ⇒ 00:07:06.039 Pranav: But end-to-end, how it should work is, essentially, it should be pulling data from Meta and Shopify, those are the two that they get. And there’s one other that is not integrated yet, it’s not in scope, but those are the two that I’m gonna be focusing on. Meta and Shopify, pulling data from there, and then…
85 00:07:06.090 ⇒ 00:07:18.590 Pranav: there are certain tables that I have format from, that I have, like, a format design, like, saved, and basically, yeah, that just goes into a channel using a cron job.
86 00:07:18.590 ⇒ 00:07:19.190 Samuel Roberts: Okay.
87 00:07:19.760 ⇒ 00:07:23.549 Samuel Roberts: Sounds pretty simple for, like, this use case.
88 00:07:24.180 ⇒ 00:07:37.070 Pranav: And I think what would be cool here is, like, if I can find… because the Slack API, from everything that I’ve heard, is just, like… it’s just, like, they’ve thought of everything, it’s designed super well.
89 00:07:37.380 ⇒ 00:07:45.079 Pranav: And so I’m just gonna think of, like, some other cool ways where they can maybe even interact with, like, the output, and then maybe…
90 00:07:45.220 ⇒ 00:07:47.270 Pranav: Like, maybe they can even, like, send a…
91 00:07:47.380 ⇒ 00:07:50.920 Pranav: send a follow-up prompt and, like, get another table out of it, or…
92 00:07:51.170 ⇒ 00:08:02.239 Pranav: something of that nature. Just like we have something exciting to show tomorrow. And we’ll just see, like, with time-wise too, right? Like, I’m not gonna start, like, pushing… pushing stuff, like, one minute before the demo, but…
93 00:08:02.240 ⇒ 00:08:02.850 Samuel Roberts: Yeah.
94 00:08:03.200 ⇒ 00:08:08.000 Pranav: Yeah, so, yeah, basically once, you get the MCPs.
95 00:08:08.480 ⇒ 00:08:16.490 Pranav: up and running in dev, then I can, like, start actually getting real data from, like, Meta and Shopify, but…
96 00:08:17.050 ⇒ 00:08:28.720 Pranav: you know, this is kind of… like, we’re definitely gonna get some feedback on the Slack reports in terms of, like, what we’re getting and what they actually want, and so I’m not too concerned if, like, we’re using sample data either for… for tomorrow.
97 00:08:29.590 ⇒ 00:08:32.380 Samuel Roberts: I’m sorry, repeat that, my son’s making noise behind me.
98 00:08:32.380 ⇒ 00:08:39.520 Pranav: Oh, all good, all good. Yeah, I was just saying, once you get, like, the MCPs, like, deployed and ready and developed.
99 00:08:39.799 ⇒ 00:08:46.260 Pranav: Into, like, the develop environment, then I can pull, like, the actual data from Meta and Shopify.
100 00:08:46.560 ⇒ 00:08:48.130 Samuel Roberts: Yeah, okay, yeah, yeah.
101 00:08:48.130 ⇒ 00:08:49.320 Pranav: Yeah. What is your plan?
102 00:08:49.320 ⇒ 00:08:51.329 Samuel Roberts: What’s looking at data, like, what’s your…
103 00:08:52.910 ⇒ 00:08:55.229 Pranav: Like, what data do I need?
104 00:08:55.560 ⇒ 00:08:58.070 Samuel Roberts: Well, I… so we have the MCP servers…
105 00:08:58.070 ⇒ 00:08:58.720 Pranav: Yeah.
106 00:08:59.810 ⇒ 00:09:04.260 Samuel Roberts: But, like, that also theoretically just gives us access to the API, because we should have the keys, necessarily.
107 00:09:04.260 ⇒ 00:09:06.050 Pranav: Yeah, exactly.
108 00:09:06.520 ⇒ 00:09:12.429 Samuel Roberts: So I… yeah, as long as I… all those keys are kind of stored and encrypted in the database, so you could really do whatever you want for the…
109 00:09:12.680 ⇒ 00:09:15.380 Samuel Roberts: this, like, the MCP or the…
110 00:09:16.630 ⇒ 00:09:20.149 Samuel Roberts: Just API, I guess. I might just be thinking out loud now.
111 00:09:22.430 ⇒ 00:09:23.220 Pranav: Yeah.
112 00:09:23.710 ⇒ 00:09:25.469 Samuel Roberts: Okay, hold on, while I’m…
113 00:09:25.760 ⇒ 00:09:31.069 Samuel Roberts: Hold on, what is going on with my computer? Where… my mouse? There we go. Okay, so I was…
114 00:09:31.670 ⇒ 00:09:36.970 Samuel Roberts: The… totally changing subjects here, just because I’m trying to figure out this deployment stuff.
115 00:09:36.970 ⇒ 00:09:37.910 Pranav: Yeah, yeah.
116 00:09:37.910 ⇒ 00:09:43.240 Samuel Roberts: How, the… the names for things?
117 00:09:43.400 ⇒ 00:09:46.309 Pranav: Yeah, it’s a little bit weird, because there’s staging and develop, right?
118 00:09:46.310 ⇒ 00:09:51.230 Samuel Roberts: Is staging… oh, yeah, so I was like, but not just that, like, Inside Railway… I…
119 00:09:51.490 ⇒ 00:09:54.039 Pranav: Like, we have API and front-end.
120 00:09:54.970 ⇒ 00:09:55.500 Pranav: Yeah.
121 00:09:55.500 ⇒ 00:09:57.940 Samuel Roberts: Is it cool if I start renaming things… now?
122 00:09:58.130 ⇒ 00:10:01.710 Pranav: Yeah, sure, if you want to. Yeah.
123 00:10:02.110 ⇒ 00:10:07.409 Samuel Roberts: I might just make it back-end, and then, like, I’m trying to try to keep the names consistent with the folder to, like, stitch…
124 00:10:07.660 ⇒ 00:10:15.189 Samuel Roberts: Shopify MCP, Stitch Claviyo MCP, Stitch Meta AdsMCP… okay, yeah, I’m just gonna do that. I’m gonna… I’m testing this out in another environment, and then I’ll push it into…
125 00:10:15.860 ⇒ 00:10:17.680 Samuel Roberts: Dev.
126 00:10:18.020 ⇒ 00:10:19.090 Pranav: Yeah, yeah, go for it.
127 00:10:19.300 ⇒ 00:10:25.640 Samuel Roberts: Okay, I just wasn’t… I… Sorry, I’m just… Hittin’… hitting a wall here.
128 00:10:25.640 ⇒ 00:10:29.140 Pranav: You’re good, you’re good. Okay. I think we’re,
129 00:10:29.340 ⇒ 00:10:34.169 Pranav: we’re making good pace on this project, especially Phase 1.
130 00:10:34.550 ⇒ 00:10:35.250 Samuel Roberts: Yeah.
131 00:10:36.060 ⇒ 00:10:41.009 Pranav: You know, like, I think by the end of next week, what we need to get done is,
132 00:10:42.770 ⇒ 00:10:44.880 Pranav: Let me look at linear again.
133 00:10:44.880 ⇒ 00:10:47.170 Samuel Roberts: Yeah, we talked about that a little bit. They have an all…
134 00:10:47.170 ⇒ 00:10:53.120 Pranav: hands on Thursday, that they were hoping to show them, like, the prototypes that they have, so… Yeah. I’m hoping that we… if we…
135 00:10:53.120 ⇒ 00:10:57.230 Samuel Roberts: you know, if I can get this deployment done today slash tomorrow now,
136 00:10:58.120 ⇒ 00:11:02.660 Samuel Roberts: That we can actually have them in, like, the staging environment, and maybe even get some other people in there to start testing it out.
137 00:11:03.360 ⇒ 00:11:09.480 Samuel Roberts: Yeah, I think we can definitely get there. Yeah, I’m not that worried about by Thursday, at this point.
138 00:11:09.850 ⇒ 00:11:10.500 Pranav: Either?
139 00:11:10.500 ⇒ 00:11:11.940 Samuel Roberts: Tomorrow, we’ll see.
140 00:11:12.170 ⇒ 00:11:21.499 Pranav: Yeah, me neither. And, like, the final thing that I can… if we need Friday, just because we have it, is just kind of onboarding Bobby and Zach,
141 00:11:21.500 ⇒ 00:11:22.240 Samuel Roberts: Yeah, yeah.
142 00:11:22.240 ⇒ 00:11:34.390 Pranav: Cloud Code. One other thing that I thought of, though, is, like, maybe we can onboard them into codecs, like, OpenAI’s, like, it’s kind of like a… it’s in the browser.
143 00:11:34.390 ⇒ 00:11:36.859 Samuel Roberts: Yeah, I hadn’t thought of that, that’s a good point.
144 00:11:36.860 ⇒ 00:11:40.780 Pranav: The good thing about that is that they don’t need to worry about, like, using localhost.
145 00:11:40.780 ⇒ 00:11:45.870 Samuel Roberts: I was… yeah, that’s one thing that worries me, so that’s a really good, good thought there.
146 00:11:45.870 ⇒ 00:11:49.899 Pranav: Yeah, so I think I’m going to first look into that, okay.
147 00:11:51.170 ⇒ 00:11:59.939 Pranav: And then if that seems to be more complex, then… I mean, I don’t think it will, so I think that is the… that is actually the best way to go about things.
148 00:11:59.940 ⇒ 00:12:00.580 Samuel Roberts: Goodbye.
149 00:12:00.900 ⇒ 00:12:06.439 Pranav: And I think they would appreciate that more, too, because then, yeah, they could easily just make a feature off of production.
150 00:12:06.440 ⇒ 00:12:07.400 Samuel Roberts: Right.
151 00:12:08.370 ⇒ 00:12:13.509 Pranav: And then it’s really easy for us to then integrate that into, like, actual usable features, so…
152 00:12:13.510 ⇒ 00:12:15.109 Samuel Roberts: Yeah, I think that makes sense, okay.
153 00:12:15.290 ⇒ 00:12:15.840 Pranav: Cool.
154 00:12:17.350 ⇒ 00:12:20.079 Samuel Roberts: Alright, well, let me, I don’t know…
155 00:12:21.030 ⇒ 00:12:23.939 Samuel Roberts: I’m gonna try to get this going, but it’s getting later than I…
156 00:12:23.940 ⇒ 00:12:33.359 Pranav: anticipated at this point, because I moved… started moving slower, like, halfway through the day. I think you’re good, yeah, I mean, I’m gonna have to drop right after this call, too.
157 00:12:34.160 ⇒ 00:12:35.710 Pranav: And so…
158 00:12:36.320 ⇒ 00:12:45.129 Pranav: Yeah, if you want to, like, start working, like, just get it deployed, like, tomorrow morning, and I could even help wherever I… wherever you need me to, so just let me know.
159 00:12:47.970 ⇒ 00:12:57.899 Pranav: And I can, like, show you how I developed… I mean, deployed the other MCP server, if it helps at all. It’s all in there too, like, the build commands, the start commands…
160 00:12:58.100 ⇒ 00:13:06.489 Samuel Roberts: Yeah, no, I was able to look at all that. I had a cursor looking at that and generating, like, a script to sync all the environment variables and then push everything.
161 00:13:06.960 ⇒ 00:13:13.149 Samuel Roberts: Okay. With the, generated URLs of those services, too, if that makes sense?
162 00:13:13.610 ⇒ 00:13:14.150 Samuel Roberts: So, like…
163 00:13:14.620 ⇒ 00:13:15.210 Pranav: Yeah.
164 00:13:15.210 ⇒ 00:13:19.559 Samuel Roberts: Yeah, just the way the other MCP server is, but now we have 3 others to worry about, so…
165 00:13:20.200 ⇒ 00:13:26.070 Samuel Roberts: Just trying to keep them in sync and not have to go to the dashboard every time we need to add something, remove something.
166 00:13:26.410 ⇒ 00:13:28.069 Pranav: Yup, yup. Okay, cool.
167 00:13:28.700 ⇒ 00:13:29.460 Samuel Roberts: Hey, Tom.
168 00:13:29.910 ⇒ 00:13:31.469 Uttam Kumaran: Hey. Hey, guys.
169 00:13:32.710 ⇒ 00:13:33.990 Uttam Kumaran: How’s everything?
170 00:13:37.790 ⇒ 00:13:40.790 Samuel Roberts: Sam’s a little under the weather, I’m… Yeah.
171 00:13:40.790 ⇒ 00:13:47.110 Pranav: My allergies, like, Utam, like, right now, like, I don’t know if you have, like, spring allergies, but, like, in Austin, it’s terrible right now.
172 00:13:47.240 ⇒ 00:13:47.750 Pranav: Like…
173 00:13:47.750 ⇒ 00:13:53.020 Uttam Kumaran: Yeah, I don’t really have allergies, like, my family gets a lot, but I’ve never really been affected.
174 00:13:53.360 ⇒ 00:13:54.250 Pranav: Yeah.
175 00:13:54.250 ⇒ 00:13:54.720 Uttam Kumaran: Yeah.
176 00:13:54.720 ⇒ 00:13:56.039 Pranav: Terrible in Austin, like…
177 00:13:56.040 ⇒ 00:13:56.960 Uttam Kumaran: That’s true.
178 00:13:56.960 ⇒ 00:13:59.399 Pranav: Massachusetts, it’s one thing, but, like, Austin…
179 00:13:59.400 ⇒ 00:14:00.979 Samuel Roberts: Especially in winter, yeah.
180 00:14:01.120 ⇒ 00:14:03.479 Pranav: And I forgot, it’s like… it’s not like…
181 00:14:04.110 ⇒ 00:14:06.919 Pranav: One good thing about winter is that, okay, there’s no pollen out.
182 00:14:06.920 ⇒ 00:14:07.800 Samuel Roberts: Right, right.
183 00:14:07.800 ⇒ 00:14:08.810 Pranav: thing in Austin.
184 00:14:09.910 ⇒ 00:14:14.199 Pranav: But, yeah, just drinking tea all day. I’m, I’m getting through it. Okay, good. Yeah, yeah.
185 00:14:15.630 ⇒ 00:14:31.479 Pranav: But, yeah, so they seemed to like the demo that, like, Sam showed, so it seems like we’re pretty aligned on, like, what they wanted for MCP servers. Okay, cool. What I was saying to Sam was that I’m also working parallelily on, like, the Slack reports. I was talking to Casey
186 00:14:31.480 ⇒ 00:14:42.589 Pranav: about, kind of, the work that he’s done there, and… Right. This is, like, from what I saw with, like, the previous agency, how they implemented it, it was, like, super bare bones, and so, like, whatever we implement is gonna be, like.
187 00:14:42.790 ⇒ 00:14:44.370 Pranav: It’s gonna be better, basically.
188 00:14:44.370 ⇒ 00:14:44.970 Uttam Kumaran: Okay.
189 00:14:44.970 ⇒ 00:14:47.480 Pranav: And so…
190 00:14:47.610 ⇒ 00:14:55.520 Pranav: what would be best would be if we can actually, like, pull the data from Meta and from Shopify, and have that
191 00:14:55.700 ⇒ 00:15:09.099 Pranav: be, like, dynamic data within the table that we then send to a Slack channel. But for tomorrow, if we can’t, like, get to the point where we can do that, like, I’ll already have, like, sample data where we’re kind of…
192 00:15:10.050 ⇒ 00:15:13.910 Pranav: putting into a table and then sending into the Slack channel.
193 00:15:14.980 ⇒ 00:15:19.420 Pranav: That’s only, like, if we just can’t, you know, we have issues with, like, deploying and stuff, and like…
194 00:15:19.580 ⇒ 00:15:23.769 Pranav: Oh, okay. I still want to be able to demo the Slack report stuff, basically.
195 00:15:24.010 ⇒ 00:15:25.330 Uttam Kumaran: Okay, cool.
196 00:15:25.510 ⇒ 00:15:26.260 Pranav: Right.
197 00:15:27.660 ⇒ 00:15:30.979 Uttam Kumaran: Yeah, so, like, I mean, in our meeting today, I just, like…
198 00:15:31.090 ⇒ 00:15:38.340 Uttam Kumaran: I think, look, we’re… just like every situation we walk into, we’re coming in because wasn’t issues.
199 00:15:38.690 ⇒ 00:15:39.350 Samuel Roberts: Right.
200 00:15:40.010 ⇒ 00:15:43.880 Uttam Kumaran: In an environment where they just ha- they just wanted to get, like, a lot of this stuff.
201 00:15:44.370 ⇒ 00:15:49.920 Uttam Kumaran: Or, like… Months ago. And so, our job is just to, like.
202 00:15:50.310 ⇒ 00:15:54.249 Uttam Kumaran: Biased in to know, like, we’re doing it the right way, but then also, like.
203 00:15:54.450 ⇒ 00:15:57.019 Uttam Kumaran: still ship, so I think we did a good job, you know?
204 00:15:57.830 ⇒ 00:15:58.390 Pranav: Cool.
205 00:15:58.880 ⇒ 00:15:59.370 Uttam Kumaran: Yeah.
206 00:15:59.370 ⇒ 00:16:09.909 Samuel Roberts: Yeah, no, they… I mean, he… Bobby even said, like, I know it sounds like they’re pushing a lot, but, like, I think they might understand that we’re going pretty quick, and we’re catching up, to where they want to be, or wanted to be, at least.
207 00:16:10.590 ⇒ 00:16:15.520 Uttam Kumaran: Yeah, yeah. And I think, you know, overall, like, our… I think our… I’m totally fine.
208 00:16:15.520 ⇒ 00:16:19.260 Samuel Roberts: And with our pace, and I knew that we were going to…
209 00:16:19.280 ⇒ 00:16:32.669 Uttam Kumaran: extend or increase. Right. I just, like… ultimately, I should have, like… and this is what we just learned, is, like, I probably should have sold a bigger budget off-rip, but I just didn’t know…
210 00:16:32.990 ⇒ 00:16:41.429 Uttam Kumaran: you know, what the whole situation was, and so… it’s okay, we’re just learning as we go, but I feel like they’ll be… they’ll be good to at least bring Casey on.
211 00:16:41.830 ⇒ 00:16:43.930 Uttam Kumaran: And we’ll be in a good spot, so…
212 00:16:44.520 ⇒ 00:16:45.080 Pranav: Cool.
213 00:16:46.120 ⇒ 00:16:53.310 Pranav: Also, Sam mentioned, too, like, on Thursday next week, they have, like, an all-hands where they would like to, like, demo some stuff, right? So…
214 00:16:53.440 ⇒ 00:16:58.949 Pranav: We feel pretty good about that, me and Sam, about, like, getting everything, like, tidy up by then.
215 00:16:58.950 ⇒ 00:16:59.480 Uttam Kumaran: Okay.
216 00:16:59.650 ⇒ 00:17:00.170 Pranav: Yeah.
217 00:17:00.170 ⇒ 00:17:05.060 Samuel Roberts: Yeah, I feel… I mean, this deployment I’m working on right now, if I can’t get this done tonight, it’ll be done tomorrow.
218 00:17:05.500 ⇒ 00:17:07.829 Samuel Roberts: At least the staging, then.
219 00:17:09.750 ⇒ 00:17:13.699 Samuel Roberts: I don’t necessarily know about getting, like, a prod all the way done, but we can, you know…
220 00:17:14.160 ⇒ 00:17:21.700 Samuel Roberts: get them in there testing it before we really… because there’s going to be things breaking when they start asking questions that I’m not asking out of the MCPs and stuff.
221 00:17:23.290 ⇒ 00:17:25.990 Samuel Roberts: But as long as the data’s flowing, I’m less worried.
222 00:17:30.920 ⇒ 00:17:31.750 Samuel Roberts: Core.
223 00:17:31.750 ⇒ 00:17:42.640 Pranav: Yeah, in terms of, like, agenda tomorrow, Utam, I kind of just, like, copied your agenda that you sent into the channel, and I just put that in as a…
224 00:17:42.640 ⇒ 00:17:51.839 Pranav: like, a linear ticket, and we’re looking pretty good on that. I think one thing that you wanted to talk about was just, like, the features from competitors on Nano Banana.
225 00:17:53.530 ⇒ 00:17:58.220 Uttam Kumaran: Yeah, I guess I… I just wanted to, like… part of these Friday calls, like.
226 00:17:58.510 ⇒ 00:18:00.419 Uttam Kumaran: I would kind of want to talk about…
227 00:18:01.180 ⇒ 00:18:07.300 Uttam Kumaran: a little bit of, like, present, and then I try to always just, like, have some topics for discussion.
228 00:18:07.500 ⇒ 00:18:12.509 Uttam Kumaran: So, like, I would… it would be great if we were able to just discuss, like.
229 00:18:12.780 ⇒ 00:18:17.059 Uttam Kumaran: You know, features that we’ve seen on other, sort of, image generation and…
230 00:18:17.130 ⇒ 00:18:33.010 Uttam Kumaran: kind of get some product feedback from them, and then, yeah, like, every week, I think we’ll just figure out, like, these guys are… it’s not gonna be so much planning, I think they’re happy with however we’re… we’re planning it, it’s just… I think every week, we’re gonna confirm, like, what features are.
231 00:18:33.090 ⇒ 00:18:37.740 Uttam Kumaran: What features are most important? And, you know, how best to…
232 00:18:37.990 ⇒ 00:18:46.999 Uttam Kumaran: how best to execute it. Additionally, like, when they request new features, we’re just gonna review, like, timeline, so we’ll think about, like, what the best way is, you know, over time to do that, so…
233 00:18:49.450 ⇒ 00:18:50.260 Uttam Kumaran: Alright.
234 00:18:52.600 ⇒ 00:18:57.269 Pranav: So, in the meeting today, did you guys go over the Gantt charts, or was it…
235 00:18:57.410 ⇒ 00:18:58.859 Pranav: It didn’t go into that much depth.
236 00:18:58.860 ⇒ 00:18:59.800 Uttam Kumaran: No, we did, we did.
237 00:18:59.800 ⇒ 00:19:10.359 Pranav: Oh, okay, cool. And so that’s where you guys kinda, like, saw, like, the… the different tasks, and basically just, like, checked off a few that maybe just didn’t need to be in scope.
238 00:19:10.570 ⇒ 00:19:15.679 Uttam Kumaran: Yeah, they, they, like, they looked at it, and they’re like, look, we have a… we have a much, like, slimmer version of…
239 00:19:16.810 ⇒ 00:19:30.729 Uttam Kumaran: the forecasting thing that we want to do, maybe get… they said get that done first, and it actually will take out, like, a lot of the cohorting stuff and some of the Python stuff. So, I think it’ll just allow us to speed up a little bit.
240 00:19:30.880 ⇒ 00:19:35.599 Uttam Kumaran: But I told them that, like, we’re gonna need a little bit more budget in order to bring on
241 00:19:37.070 ⇒ 00:19:50.869 Uttam Kumaran: just Casey, and so, yeah, and then they’re already looking forward to, like, working with us for a longer time, so I basically said, like, look, I want to roll on Hernav to just be kind of, like, dedicated here, and then have Casey support, and then…
242 00:19:51.130 ⇒ 00:19:55.100 Uttam Kumaran: I think eventually, like, we’ll… I’ll probably,
243 00:19:56.130 ⇒ 00:20:03.629 Uttam Kumaran: once Gabe joins, I’ll probably loop in Gabe to also start to get involved here, kind of, like, take me out of the loop a little bit, but
244 00:20:03.930 ⇒ 00:20:09.659 Uttam Kumaran: Yeah, so I’m gonna make some of the changes to the Gantt chart, probably later today.
245 00:20:10.060 ⇒ 00:20:12.899 Uttam Kumaran: And then… Cool. Yeah, we should be in a good spot.
246 00:20:13.700 ⇒ 00:20:16.490 Pranav: Gotcha. And you were saying, like, Gabe might step in as, like, CSO?
247 00:20:16.970 ⇒ 00:20:22.279 Uttam Kumaran: I’m not sure yet, like, cause right now we have a couple projects where we kind of need, like.
248 00:20:22.530 ⇒ 00:20:24.980 Uttam Kumaran: Almost like a product person.
249 00:20:24.980 ⇒ 00:20:27.910 Pranav: So I don’t see it necessarily as, like.
250 00:20:28.480 ⇒ 00:20:35.440 Uttam Kumaran: CSO, it’s… I’m not sure yet, exactly. Because a lot of our data projects, we don’t have, like, a product manager.
251 00:20:37.020 ⇒ 00:20:37.509 Samuel Roberts: the different…
252 00:20:37.760 ⇒ 00:20:47.979 Uttam Kumaran: So, yeah, I don’t know what the final, like, structure is gonna be. I just think Gabe, overall, is gonna… is starting to do a lot of, like, basically, like, AI feature for, like, product work, so…
253 00:20:47.980 ⇒ 00:20:48.330 Pranav: Okay.
254 00:20:48.330 ⇒ 00:20:54.949 Uttam Kumaran: clients where that’s, like, needed, maybe he fills in. But yeah, I don’t… not 100% sure yet.
255 00:20:55.970 ⇒ 00:20:57.760 Pranav: Yeah, I think,
256 00:20:58.070 ⇒ 00:21:04.590 Pranav: especially in the initial phases, like, when we’re doing the scoping, when we’re doing the estimates, the Gantt charts, like.
257 00:21:04.930 ⇒ 00:21:24.770 Pranav: sometimes conceptualizing what that product looks like is, like, super useful, and then after that, it’s just really just, like, implementation. Yeah. Like, yeah, just, like, in the first couple weeks, like, maybe even just, like, next week, just having Gabe around to kind of, like, make sure we’re thinking about things in the right way, and, like, we’ve…
258 00:21:24.830 ⇒ 00:21:30.309 Pranav: covered all of our bases in terms of, like, what the product needs to look like. That could… that could be helpful, yeah. But I don’t know if it needs.
259 00:21:30.310 ⇒ 00:21:30.720 Uttam Kumaran: It’d be like.
260 00:21:30.720 ⇒ 00:21:32.759 Pranav: Somebody for the entirety of a project.
261 00:21:32.990 ⇒ 00:21:34.140 Uttam Kumaran: Okay, okay, okay.
262 00:21:34.280 ⇒ 00:21:36.999 Uttam Kumaran: Yeah, I also agree. I don’t know whether…
263 00:21:37.280 ⇒ 00:21:49.760 Uttam Kumaran: it, like, if there’s consistent, like, product development work, but, like, for stuff like this, like, it would have been nice for someone to have done, like, the design before handing it to you guys. Totally. You know, and, like, a couple things like that, so…
264 00:21:51.270 ⇒ 00:22:02.529 Uttam Kumaran: Yeah, I just think, again, like, as we start to get more clients that are doing, like, sort of application development, we’re gonna bake in… I’m just learning every time we do these, so we’re gonna bake in time for, like, product
265 00:22:02.690 ⇒ 00:22:04.120 Uttam Kumaran: Product management.
266 00:22:04.270 ⇒ 00:22:15.570 Uttam Kumaran: And then, so someone will come in, scope out flows, and then basically get those approved, and then hand it off, and we’ll just budget for that. Because otherwise, like, some… we’ll just end up having to make all these product decisions, and…
267 00:22:15.790 ⇒ 00:22:19.079 Uttam Kumaran: I’d rather them be, like, much more well thought out, you know, so…
268 00:22:19.590 ⇒ 00:22:20.340 Pranav: Yeah.
269 00:22:20.420 ⇒ 00:22:23.759 Samuel Roberts: Also, one thing that I noticed with this week was, like.
270 00:22:23.760 ⇒ 00:22:28.980 Pranav: I had to spend, like, a good amount of time on, like, estimating and, like, Gantt charts, and just.
271 00:22:28.980 ⇒ 00:22:29.360 Uttam Kumaran: Yeah.
272 00:22:29.360 ⇒ 00:22:34.669 Pranav: conversations and stuff, and I know that’s not gonna be the case, like, going forward, but I wonder if, like.
273 00:22:34.800 ⇒ 00:22:38.140 Pranav: That’s just, like, to be expected, or if we want to kind of…
274 00:22:38.420 ⇒ 00:22:45.999 Pranav: I… I feel like that’s totally fine, and it’s probably… it’s gonna be… it’s gonna have to happen regardless.
275 00:22:46.000 ⇒ 00:22:49.920 Uttam Kumaran: A lot of… well, a lot of the estimation, like, I… the PRDs.
276 00:22:50.150 ⇒ 00:22:57.820 Uttam Kumaran: like, kind of have been done, so a lot of the estimation thing is just waiting for us to kind of, like, get them done. It’s not gonna be every week where we, like, scope out, like.
277 00:22:57.920 ⇒ 00:22:59.510 Uttam Kumaran: Two months’ worth of work.
278 00:22:59.610 ⇒ 00:23:00.870 Pranav: Yeah, exactly.
279 00:23:00.870 ⇒ 00:23:02.120 Uttam Kumaran: Yeah. So more of, like.
280 00:23:02.480 ⇒ 00:23:03.190 Samuel Roberts: Understood.
281 00:23:03.190 ⇒ 00:23:12.640 Uttam Kumaran: Yeah, exactly. Like, I think when we go into new contracts, there will be, like, a week of sort of planning, but usually it’ll just be, like, adjustments.
282 00:23:12.700 ⇒ 00:23:23.749 Uttam Kumaran: So, the thing about all of our projects is, like, we just didn’t get a chance to do a lot of that for every project that we signed before the break. So now it’s… we’re sort of, like, playing catch-up.
283 00:23:23.850 ⇒ 00:23:30.989 Uttam Kumaran: So now that we have EPs everywhere, everybody’s sort of, like, raising the floor, so at least every client has one of them. I mean…
284 00:23:31.020 ⇒ 00:23:45.680 Uttam Kumaran: basically, as we start, every new client will, ideally, as part of the, like, SOW process, depending on, like, how fast we can get the whole flow, like, maybe we develop the first version of the Gantt as part of the SOW.
285 00:23:46.090 ⇒ 00:23:52.010 Uttam Kumaran: Even before the project starts, and maybe, like, sales has to do that, or sales and tech.
286 00:23:52.330 ⇒ 00:23:57.230 Uttam Kumaran: But ideally, first week of the project… first week of the project will be more…
287 00:23:57.370 ⇒ 00:24:06.250 Uttam Kumaran: like, planning than development work. Right now, we just got jammed, because we signed all these clients, and we just, like, had to figure shit out before break.
288 00:24:06.420 ⇒ 00:24:12.990 Uttam Kumaran: But going forward, yeah, we’re gonna start… we know, like, what we expect per project, and so it’ll start off
289 00:24:13.360 ⇒ 00:24:17.619 Uttam Kumaran: Like, we’ll budget a week in the beginning just for planning and discovery, mainly.
290 00:24:18.140 ⇒ 00:24:23.340 Pranav: Yeah, I think that makes a lot of sense, and also just for the client to be aware of that, too. You know, like.
291 00:24:23.500 ⇒ 00:24:27.449 Uttam Kumaran: Yeah, we’ll bake… exactly, we’ll just tell them that that’s how we operate, yeah.
292 00:24:27.450 ⇒ 00:24:36.170 Pranav: Yeah, they’re not gonna expect to see, like, a ton of features that first week, because I feel like this week, like, yeah, Sam kind of, like, got to focus on the MCP stuff, but…
293 00:24:36.330 ⇒ 00:24:43.000 Pranav: I couldn’t focus a ton on, like, coding until, like, yesterday and today.
294 00:24:43.000 ⇒ 00:24:43.530 Uttam Kumaran: Yeah.
295 00:24:43.530 ⇒ 00:24:47.430 Pranav: Yesterday. So, like, just so, like, they’re aware, you know? It’s just, like, keeping them in.
296 00:24:47.430 ⇒ 00:24:48.020 Uttam Kumaran: Yeah.
297 00:24:48.020 ⇒ 00:24:51.740 Pranav: Most of this stuff is, like, it only is gonna work in our… work in our favor.
298 00:24:51.740 ⇒ 00:25:07.470 Uttam Kumaran: And this is, like, so this is mostly for, like, most of the people in the company, so this is everybody’s first time writing SOWs, writing decks, or putting GANs together. Yeah. So the next time will be 50% faster, and then everybody will reach, like… it will never take this long.
299 00:25:07.770 ⇒ 00:25:09.190 Pranav: Going forward.
300 00:25:09.190 ⇒ 00:25:14.410 Uttam Kumaran: Not only to write it, but to get feedback, and so, again, like, we paid the…
301 00:25:14.890 ⇒ 00:25:19.329 Uttam Kumaran: We just have to pay the tool on this one to figure out, like, how to scale this whole model, so…
302 00:25:19.590 ⇒ 00:25:21.510 Pranav: Yeah, that’s a good one. Yeah, it probably.
303 00:25:21.510 ⇒ 00:25:30.809 Uttam Kumaran: And again, like, we were never at risk of losing this client. It was just a risk of, like, I know we… it’s gonna take a lot more effort, and so…
304 00:25:31.030 ⇒ 00:25:33.789 Uttam Kumaran: But also, again, like, as we’re growing.
305 00:25:33.970 ⇒ 00:25:52.179 Uttam Kumaran: we’re… sales, like, we’re raising our minimum, meaning we’re not gonna take… we’re, like… I’m pushing us to stop taking clients, like, less than 15 grand a month, because it’s hard for us to staff and do a quality job with less than that, and I don’t want to take work where it’s, like.
306 00:25:52.700 ⇒ 00:26:11.819 Uttam Kumaran: where it’s, like, less money than that. And so, typically, what we’re doing is I’m pushing sales to just, like, to kind of get to that level, because then we can appropriately staff and make sure we get all the work done. But, like, again, that’s been a change that’s happened over the last, like, 4 months since our minimums are kind of going up.
307 00:26:12.050 ⇒ 00:26:17.050 Uttam Kumaran: So, again, we’re still, we’re still dealing with some clients where this just wasn’t the case.
308 00:26:17.510 ⇒ 00:26:23.609 Uttam Kumaran: When we signed him. But Lilo’s a good example where, like, I tried to budget higher.
309 00:26:23.750 ⇒ 00:26:29.009 Uttam Kumaran: We should have went even higher, and so that’s just what we’re gonna continue to do, is, like, raise prices.
310 00:26:29.280 ⇒ 00:26:40.669 Uttam Kumaran: It’s just sort of, like, when demand starts to continue to get high, our job should just be to raise prices, because ultimately, I would rather have less clients, they all pay, like, a lot than have, like.
311 00:26:40.980 ⇒ 00:26:43.200 Uttam Kumaran: 100, you know?
312 00:26:43.570 ⇒ 00:26:46.779 Samuel Roberts: Tons of little clients, it’s a ton of context switching, and lots of… yeah, it’s a…
313 00:26:46.890 ⇒ 00:26:50.389 Uttam Kumaran: Yeah, we’re a small crew, we’re not a huge crew, right?
314 00:26:50.390 ⇒ 00:26:50.950 Samuel Roberts: Exactly.
315 00:26:50.950 ⇒ 00:26:56.209 Uttam Kumaran: A lot of folks at Brainforge are on 3 to 4 clients, which is not really common.
316 00:26:56.320 ⇒ 00:27:02.599 Uttam Kumaran: So we need to sort of, like, fix one problem at a time. But, like, applying, like, Element, applying, like, CES,
317 00:27:02.820 ⇒ 00:27:08.170 Uttam Kumaran: a client, like ABC, they have a ton of growth.
318 00:27:08.210 ⇒ 00:27:10.349 Samuel Roberts: And we’re gonna probably find…
319 00:27:10.350 ⇒ 00:27:18.320 Uttam Kumaran: 3 more Lilos within all those, you know? So, I would… it’s like a mix of, like, expanding…
320 00:27:19.310 ⇒ 00:27:24.040 Uttam Kumaran: Expanding clients where there is upward mobility, and then also just, like.
321 00:27:24.160 ⇒ 00:27:27.379 Uttam Kumaran: like, raising our minimums. Like, we don’t work with a lot of, like.
322 00:27:27.510 ⇒ 00:27:29.910 Uttam Kumaran: Startups, or people where they’re like.
323 00:27:30.030 ⇒ 00:27:35.349 Uttam Kumaran: Because the problem is when you work with clients that are, like, 2, 3, 4, 5K, they’re very stingy.
324 00:27:35.850 ⇒ 00:27:45.280 Uttam Kumaran: and everything’s a discussion, versus, like, Lilo, like, you could tell, they’re like, dude, whatever, just, like, get… I don’t care how… just, like, tell us how much it is, but, like, can you get it done by, like, next month?
325 00:27:45.630 ⇒ 00:27:49.220 Uttam Kumaran: You know, and so there’s just a different vibe.
326 00:27:49.560 ⇒ 00:27:50.370 Samuel Roberts: Definitely.
327 00:27:56.540 ⇒ 00:27:58.320 Samuel Roberts: Cool.
328 00:28:00.640 ⇒ 00:28:01.739 Samuel Roberts: Traveling over here.
329 00:28:04.140 ⇒ 00:28:13.569 Uttam Kumaran: So I guess, like, we can probably start looping in Casey as early as tomorrow, and then… I guess maybe come Monday, we can…
330 00:28:13.690 ⇒ 00:28:15.249 Uttam Kumaran: Plan a bit more.
331 00:28:15.690 ⇒ 00:28:20.700 Uttam Kumaran: You know, like, what we want to do for the next… You know, 2 weeks.
332 00:28:20.910 ⇒ 00:28:21.300 Samuel Roberts: Yeah.
333 00:28:21.380 ⇒ 00:28:22.449 Uttam Kumaran: Cool.
334 00:28:22.710 ⇒ 00:28:23.980 Uttam Kumaran: Yeah, so…
335 00:28:24.650 ⇒ 00:28:28.670 Pranav: Tomorrow, I can start sending him, like, resources, I can send him the linear, okay.
336 00:28:29.500 ⇒ 00:28:35.639 Pranav: And then, basically, the GitHub, railway, and then, just give them the staging URL.
337 00:28:36.200 ⇒ 00:28:36.910 Uttam Kumaran: Okay.
338 00:28:37.470 ⇒ 00:28:38.260 Uttam Kumaran: Great.
339 00:28:40.900 ⇒ 00:28:47.359 Uttam Kumaran: And then also, Pranav, I wanted to loop you in to… our, ABC client.
340 00:28:47.610 ⇒ 00:28:48.560 Pranav: Nope.
341 00:28:48.940 ⇒ 00:28:53.009 Uttam Kumaran: Yeah, we can think about, Sam, like, where to plug.
342 00:28:53.180 ⇒ 00:28:54.450 Uttam Kumaran: Bernavin.
343 00:28:56.540 ⇒ 00:29:08.019 Uttam Kumaran: But, yeah, I think I’ll maybe leave that to you on, like, where everybody’s skill set is, but we are kind of going through, like, a little bit of a migration from NA to N, and we are also trying to ship some new features.
344 00:29:08.340 ⇒ 00:29:16.659 Uttam Kumaran: I just think this, like, next phase of, like, figuring out evals and accuracy is gonna be a bit hard, and I don’t think we’ve ever done it, like, super perfect, Sam.
345 00:29:16.920 ⇒ 00:29:17.290 Samuel Roberts: No.
346 00:29:17.290 ⇒ 00:29:21.980 Uttam Kumaran: So I do want, like, all four of the team.
347 00:29:22.440 ⇒ 00:29:25.799 Uttam Kumaran: To, like, kind of think about how to do, like, evals and…
348 00:29:27.150 ⇒ 00:29:32.699 Uttam Kumaran: Ragged, like, actually super well, because it’s going to be a skill set that we’ll need for more other clients.
349 00:29:33.070 ⇒ 00:29:34.230 Uttam Kumaran: So…
350 00:29:36.000 ⇒ 00:29:46.919 Samuel Roberts: Yeah, the evals actually might make sense, because, like, Mustafa’s kind of, like, doing that and, like, the migration and some of the testing. Not the migration, like, the architecture stuff for, like, figuring out the new agents and sub-agents.
351 00:29:47.420 ⇒ 00:29:47.780 Samuel Roberts: Yeah.
352 00:29:47.990 ⇒ 00:29:52.279 Samuel Roberts: paralyze that a little bit, more than just, like… yeah, I think that that could work really well.
353 00:29:53.340 ⇒ 00:30:00.659 Uttam Kumaran: I just don’t think we ever, like, nailed it on, like, having evals and, kind of, like, setting up alerting when, like.
354 00:30:00.960 ⇒ 00:30:12.429 Samuel Roberts: Accuracy goes down, or something goes down, you know, so… Yeah, no, I think… I think getting off N8N is gonna help that, because we can see things… we can be logging things a lot better than it was doing.
355 00:30:13.290 ⇒ 00:30:20.490 Samuel Roberts: And then turning that into the evals is the other side of it that we only sort of just started. Like, the thumbs up, thumbs down stuff, actually analyzing that.
356 00:30:21.630 ⇒ 00:30:22.050 Uttam Kumaran: Yeah.
357 00:30:22.400 ⇒ 00:30:23.230 Samuel Roberts: Yeah.
358 00:30:24.120 ⇒ 00:30:24.830 Samuel Roberts: Hmm.
359 00:30:25.270 ⇒ 00:30:33.210 Uttam Kumaran: Yeah, I mean, I think you and… you two are, like, kind of probably the most senior on this team, so what I kind of want you guys to start to see is, like.
360 00:30:33.430 ⇒ 00:30:37.569 Uttam Kumaran: What are the things that we need to do across every client when we deal with
361 00:30:37.770 ⇒ 00:30:44.349 Uttam Kumaran: like, AI environments, and then… That way, we kind of create playbooks, because ultimately, like.
362 00:30:44.540 ⇒ 00:30:57.749 Uttam Kumaran: y’all’s time is best leveraged going into the next new client, and sort of, like, figuring it out, and creating reasonable playbooks. Like, I’m gonna go get, you know, engineers that can, you know, kind of go
363 00:30:57.960 ⇒ 00:31:01.059 Uttam Kumaran: Under y’all, and that you guys can kind of direct?
364 00:31:01.520 ⇒ 00:31:03.499 Uttam Kumaran: So, like, figuring out, like.
365 00:31:03.630 ⇒ 00:31:09.620 Uttam Kumaran: how to do effective evals. We’re gonna probably have to set up a very similar thing for… Or,
366 00:31:09.940 ⇒ 00:31:15.180 Uttam Kumaran: For internally, and for, for Lilo as well, so…
367 00:31:15.180 ⇒ 00:31:15.780 Samuel Roberts: Great.
368 00:31:18.750 ⇒ 00:31:19.410 Samuel Roberts: Yeah.
369 00:31:22.610 ⇒ 00:31:24.819 Pranav: You guys think that would be starting next week?
370 00:31:27.330 ⇒ 00:31:37.580 Uttam Kumaran: Yeah, I think we’ll probably end up starting it next week. I think still, like, probably end up spending most of your time on Lilo stuff. Yeah. But I do think that…
371 00:31:37.780 ⇒ 00:31:43.850 Uttam Kumaran: Yeah, I think however we’re gonna organize… I’m gonna… this week was actually, like.
372 00:31:44.240 ⇒ 00:31:48.600 Uttam Kumaran: Most of my time just went to, like, kind of figuring out this new, like, company structure, but, like, default.
373 00:31:48.600 ⇒ 00:31:48.930 Pranav: Right.
374 00:31:48.930 ⇒ 00:31:56.300 Uttam Kumaran: is, like, pretty off my hands. Magic Spoon, too, is, like, so… trying to get some more free time that’ll go just to, like.
375 00:31:56.610 ⇒ 00:32:02.109 Uttam Kumaran: Probably more… mostly… Abc, and then I think…
376 00:32:02.230 ⇒ 00:32:06.270 Uttam Kumaran: As long as we just keep, like, a good pace on Lilo, like, we’ll be totally fine.
377 00:32:06.520 ⇒ 00:32:07.650 Uttam Kumaran: So… Yeah.
378 00:32:08.410 ⇒ 00:32:09.930 Uttam Kumaran: Yeah.
379 00:32:12.700 ⇒ 00:32:14.610 Uttam Kumaran: But yeah, probably starting next week.
380 00:32:15.660 ⇒ 00:32:16.890 Pranav: Cool, sounds good.
381 00:32:20.270 ⇒ 00:32:21.780 Pranav: You still feeling pizza?
382 00:32:23.510 ⇒ 00:32:24.830 Uttam Kumaran: Yeah, dude, how do you feel?
383 00:32:24.990 ⇒ 00:32:26.190 Pranav: I feel good, I feel very good.
384 00:32:26.190 ⇒ 00:32:26.759 Uttam Kumaran: Is it?
385 00:32:27.150 ⇒ 00:32:35.360 Pranav: Shaishi’s gonna meet us there, okay. It’s, like, 15 minutes from me. So, yeah, I’ll probably head out soon.
386 00:32:36.490 ⇒ 00:32:45.410 Uttam Kumaran: Okay, sick, yeah, I’m just taking the dog for a walk, so I’m down. Yeah, I tried, I… I was in, like, maybe, like, 8 hours of meetings today, and I tried to, like, do some more work.
387 00:32:45.670 ⇒ 00:32:46.370 Samuel Roberts: And I was like.
388 00:32:46.370 ⇒ 00:32:54.560 Uttam Kumaran: I can’t even, like, open my eyes. So I looked at some Figma… I looked at some Figma marketing things and left some comments, and I was like.
389 00:32:54.980 ⇒ 00:32:57.570 Uttam Kumaran: I gotta go take a break, come back.
390 00:32:58.010 ⇒ 00:33:00.060 Uttam Kumaran: Come back to the night shift.
391 00:33:01.960 ⇒ 00:33:05.950 Uttam Kumaran: Yeah, yeah, I’ll be there, I’ll be there. I live, like, 5 minutes from there.
392 00:33:06.130 ⇒ 00:33:07.680 Pranav: Oh, perfect, perfect.
393 00:33:07.680 ⇒ 00:33:12.670 Uttam Kumaran: Yeah, so I’ll be there. We should take a… we’ll take a picture and try to send it to the company.
394 00:33:12.940 ⇒ 00:33:15.289 Pranav: Yeah, yeah. Try to bait Sam to move down here.
395 00:33:15.440 ⇒ 00:33:16.390 Samuel Roberts: Yeah.
396 00:33:17.820 ⇒ 00:33:19.014 Uttam Kumaran: Yes!
397 00:33:20.830 ⇒ 00:33:25.779 Uttam Kumaran: I have a good feeling some more people. I’m trying to get Amber to move down here, she’s in between apartments.
398 00:33:26.270 ⇒ 00:33:27.009 Pranav: Oh, okay.
399 00:33:27.150 ⇒ 00:33:28.199 Uttam Kumaran: Where is she?
400 00:33:28.530 ⇒ 00:33:29.730 Pranav: Is she in New York?
401 00:33:30.740 ⇒ 00:33:31.200 Uttam Kumaran: Yeah.
402 00:33:31.870 ⇒ 00:33:32.200 Pranav: Okay.
403 00:33:32.200 ⇒ 00:33:34.150 Uttam Kumaran: She’s, she’s in LA.
404 00:33:34.540 ⇒ 00:33:35.690 Pranav: Oh, LA, okay.
405 00:33:36.290 ⇒ 00:33:46.039 Uttam Kumaran: Yeah, and dude, you know, ABC is actually… is actually here, near my house. It’s off 290. So, if you start working on Andy, we can go into the office and see them.
406 00:33:46.040 ⇒ 00:33:46.760 Pranav: Reset, yeah.
407 00:33:46.760 ⇒ 00:33:52.089 Uttam Kumaran: You’ll see, like, after you start working on them, you’ll see… you’ll see the trucks everywhere around town.
408 00:33:52.220 ⇒ 00:33:53.149 Uttam Kumaran: It’s like…
409 00:33:53.150 ⇒ 00:33:54.060 Pranav: Oh, yeah.
410 00:33:54.060 ⇒ 00:33:57.740 Uttam Kumaran: I didn’t notice it until I started working for them, and then now, like.
411 00:33:58.190 ⇒ 00:34:01.279 Uttam Kumaran: You’ll see, like, I saw the truck, I see the trucks everywhere.
412 00:34:01.500 ⇒ 00:34:06.149 Uttam Kumaran: They’re a huge company here. And you’ll see all the ABC billboards on 35.
413 00:34:06.560 ⇒ 00:34:07.690 Uttam Kumaran: Like.
414 00:34:07.690 ⇒ 00:34:08.440 Pranav: Hell yeah.
415 00:34:08.659 ⇒ 00:34:12.459 Uttam Kumaran: Yeah, all over 35, they, like, buy hella big, big billboards.
416 00:34:13.480 ⇒ 00:34:14.469 Pranav: Gotcha, gotcha.
417 00:34:21.110 ⇒ 00:34:23.569 Samuel Roberts: Excuse me, alright, well, I probably need to run at this point.
418 00:34:24.030 ⇒ 00:34:25.040 Uttam Kumaran: Okay, alright.
419 00:34:25.360 ⇒ 00:34:28.269 Uttam Kumaran: Thanks, guys. Sorry, Pranav, I’ll see you in the… in just a sec, then.
420 00:34:28.270 ⇒ 00:34:28.810 Samuel Roberts: Yeah.
421 00:34:28.810 ⇒ 00:34:29.590 Pranav: Yeah, see you guys.
422 00:34:29.590 ⇒ 00:34:30.100 Samuel Roberts: Bye, guys.
423 00:34:30.100 ⇒ 00:34:32.459 Uttam Kumaran: Oh, okay. Alright, see you soon. Thank you. Bye.