Meeting Title: Eden Data Integration Technical Sync Date: 2026-04-13 Meeting participants: Pranav Narahari, Uttam Kumaran


WEBVTT

1 00:02:03.790 00:02:04.810 Uttam Kumaran: Hey, dude.

2 00:02:07.250 00:02:08.810 Pranav Narahari: Hello, hello!

3 00:02:10.620 00:02:19.720 Uttam Kumaran: Okay, so I got your… so basically today, just wanted to go through, like, that… the ChatGPT thing you sent me, or you want me to go through the larger doc as well?

4 00:02:21.430 00:02:26.039 Pranav Narahari: I think,

5 00:02:26.180 00:02:35.770 Pranav Narahari: Yeah, so the ChatGPT thing was kind of just, like, a starting point for really refining, like, the technical implementation for the…

6 00:02:36.450 00:02:43.320 Pranav Narahari: The service account, which is supposed to give us the… entirety of Eden’s organization data.

7 00:02:43.460 00:02:55.189 Pranav Narahari: Yeah. Specifically for Google, right? And then also, it includes the same… we’re also looking to do the same thing with Slack, right? So I think that ChatGPT thread gave some type of…

8 00:02:55.300 00:03:02.439 Pranav Narahari: like, technical approach that I wasn’t fully convinced on, because it was kind of trying to have us build, like, a bigger…

9 00:03:02.730 00:03:04.000 Pranav Narahari: a more…

10 00:03:04.330 00:03:11.199 Pranav Narahari: heavy system, where, like, we would have to have… do, like, some type of level of storage and stuff, so… yeah, but we can use that as, like, a starting point.

11 00:03:16.000 00:03:18.550 Uttam Kumaran: Okay, so I… yeah, go ahead.

12 00:03:18.860 00:03:32.220 Pranav Narahari: I was just so… what I kind of wanted Sam to work on this morning was, yeah, given that thread, how does he, like… I think it brought up some concerns, right? Because I basically gave it the context of our current project plan.

13 00:03:32.240 00:03:41.719 Pranav Narahari: our current technical approach, and then also, in that beginning prompt, I gave it kind of my top-of-mind ideas of how we would go about getting

14 00:03:41.810 00:03:51.389 Pranav Narahari: the entirety of the context of Eden’s… Eden’s Google data, as well as the Slack data.

15 00:03:51.610 00:04:03.249 Pranav Narahari: And then it gave… it was… it basically gave some technical approach. I just basically wanted Sam to refine the technical approach, so we felt like this was a good step going forward. But then on top of that also.

16 00:04:03.560 00:04:10.950 Pranav Narahari: there was… what I read in the chat, too, was that there was various approaches based on how…

17 00:04:12.060 00:04:27.009 Pranav Narahari: based on what we are exactly looking for. And that’s important here, because we don’t have an exact scope of work for Eden, and that’s by design. Danny kind of wants to underst… like, he kind of wants to…

18 00:04:27.170 00:04:29.479 Pranav Narahari: Direct us as we go.

19 00:04:29.760 00:04:49.219 Pranav Narahari: So for this meeting today, I wanted to get… kind of… come up with, like, hey, one, two, or two to three different approaches. What is the technical complexity, so then I can roll that to Danny and give him, like, some realistic, hey, we could get this done by this time. Here’s the…

20 00:04:49.670 00:04:59.719 Pranav Narahari: pros and cons by a different approach, and then we give our recommendation of, like, hey, let’s do this, because for your, for your usage of the command center, this should be good enough.

21 00:04:59.950 00:05:01.479 Pranav Narahari: Does that… does that make sense?

22 00:05:02.210 00:05:13.299 Uttam Kumaran: Yeah, so I guess, that’s fine. I think, like, let me give you a sense of how… so we’re… I’m doing this for Brainforge right now. So, I’m using Polytomic to pull in all of our Slack messages.

23 00:05:13.410 00:05:31.549 Uttam Kumaran: I’ve actually just set up my own Google Calendar sync that pulls all of the Google Calendar events that anyone at Brainforge is doing, so I can start to look at, like, how many people are in meetings. So, I have… there’s a process to do the Google piece. Slack, I’m using Polyatomic.

24 00:05:31.690 00:05:38.109 Uttam Kumaran: you know, you can either use Polyatomic to do… I would… that’s probably what I would suggest, or you could write your own

25 00:05:38.260 00:05:42.550 Uttam Kumaran: script to bring in those messages, but the polyatomic thing seems like it’s working well.

26 00:05:42.760 00:05:45.940 Uttam Kumaran: You actually don’t need to use…

27 00:05:46.230 00:05:48.399 Uttam Kumaran: I mean, this is where it’s like, I kind of…

28 00:05:48.680 00:05:51.890 Uttam Kumaran: I think this is where, like, I would push you to do two things. One is, like.

29 00:05:52.760 00:05:57.849 Uttam Kumaran: Have you talked into… have you spoke with Awash at all about, like, any of the data ingestion?

30 00:05:58.160 00:06:04.690 Uttam Kumaran: related items on… on… on this? Because this seems still, like, mostly data ingestion.

31 00:06:05.140 00:06:08.020 Uttam Kumaran: Like, none of this is, like, AI stuff so far.

32 00:06:08.540 00:06:20.420 Pranav Narahari: Yeah, so when we were first creating this project plan, I looped in, like, Awash and Sam, kind of gave them the, like, the same context I just gave you about, like, hey, this is what we’re trying to look… that’s just what we’re trying to build from a product standpoint.

33 00:06:20.470 00:06:34.140 Pranav Narahari: And then I was like, hey Sam, hey, Awash, you guys kind of discuss and figure out where both of you guys fit in on this. Initially, Sam had scoped out this whole, like, data warehousing,

34 00:06:34.240 00:06:49.010 Pranav Narahari: part of this project, and that’s where it made sense for Awash to step in. Once we started talking about a GWS CLI, Awash felt like, yeah, Sam, this is all going to be really AI work, there’s not really much for Awash to do here.

35 00:06:49.110 00:06:57.690 Pranav Narahari: Okay. So, yeah, Sam kind of just, like, took that over, I guess. Like, he didn’t, as part of the technical approach, as you’ve probably read, like.

36 00:06:57.790 00:07:03.759 Pranav Narahari: polyatomic’s not mentioned, so now it’s… technical approach is changing again, maybe it makes sense to loop OH back in.

37 00:07:05.530 00:07:13.069 Uttam Kumaran: Yeah, so… Yeah, this is where, like, this signal store piece, I would…

38 00:07:13.380 00:07:23.439 Uttam Kumaran: I think that’s gonna be all SAM, like, I’m not 100% sure, but in terms of getting the data somewhere, like, you have two approaches. I think…

39 00:07:23.670 00:07:34.660 Uttam Kumaran: like, if you just… for example, if you’re like, hey, I just want to see, like, today’s information, you can pull a lot of that through the CLI, but if you want to pull, like.

40 00:07:35.390 00:07:45.690 Uttam Kumaran: everything for all time, and sort of show, like, a dashboard, you would… you need to sort of have that in a data store. So I think my understanding of the project was, like.

41 00:07:46.080 00:07:52.920 Uttam Kumaran: that he wants, like, a report every day of, sort of, like, what happened in the previous day. I guess, is there a need for, like.

42 00:07:53.560 00:07:56.660 Uttam Kumaran: Looking at, like, all of the data of all time.

43 00:07:56.780 00:08:01.960 Uttam Kumaran: Because, like, this theme discovery piece is, like, It seems like they

44 00:08:02.280 00:08:04.399 Uttam Kumaran: Look at a bunch of data over time, right?

45 00:08:05.970 00:08:10.109 Pranav Narahari: Yeah, so there’s two ways that we can go about doing that, and I…

46 00:08:10.470 00:08:25.080 Pranav Narahari: this is how I thought about doing it. Every week we’re going to be generating a report, and so the data that we’re going to be looking at is for that week. Now, what are we going to measure that against? It’s going to be last week’s report.

47 00:08:25.430 00:08:32.379 Pranav Narahari: So, that will be saved in, in, like, some type of database.

48 00:08:33.020 00:08:48.479 Pranav Narahari: But we don’t need to save… we don’t need to go then and query all the data from previous weeks. We’re just going to be using that report as the… the truth that we’re querying against. So, in terms of, like, assessing trends, it’s not going to be like, hey, let’s look at

49 00:08:48.660 00:08:56.380 Pranav Narahari: for this week, we’re gonna then compare that to 2 months prior, or anything like that. We’re not gonna take that raw data, it’s just gonna be based off that report.

50 00:08:57.000 00:08:57.650 Uttam Kumaran: Okay.

51 00:08:58.480 00:09:13.990 Pranav Narahari: So that’s where I felt like, yeah, the CLI, we could maybe do that. If we can do it for a day, probably for, you know, just, it’s not… it’s not gonna be, like, in a chat interface for generating that report, it’s just gonna be on a schedule, so even if it takes a little bit longer for the…

52 00:09:14.240 00:09:23.789 Pranav Narahari: the… the GWS CLI command to run for, like, a… for… to getting a week’s data, that’s totally fine, because it’s just getting generated once a week.

53 00:09:25.260 00:09:29.620 Uttam Kumaran: Okay, and then do we have… did we get, like, service account from them yet, or…

54 00:09:29.620 00:09:30.270 Pranav Narahari: Yeah.

55 00:09:30.680 00:09:44.030 Pranav Narahari: Yeah, we have a service account, we tested it that it has DWD. The issue that we found, though, is that using the service account with the GWS CLI

56 00:09:44.230 00:09:47.780 Pranav Narahari: is causing some issues. And this is like a…

57 00:09:47.960 00:09:57.860 Pranav Narahari: you know, this is, like, a public issue. We saw that there’s, like, a PR created to kind of patch this that just has been stale. Essentially, like, with…

58 00:09:58.700 00:10:05.500 Pranav Narahari: with how DWD works, from my understanding, for a service account, is that you need to be able to

59 00:10:05.860 00:10:09.099 Pranav Narahari: The idea is that you alias as a certain individual.

60 00:10:09.430 00:10:13.130 Pranav Narahari: And that alias portion of things is not…

61 00:10:13.240 00:10:16.090 Pranav Narahari: functioning properly with the GWS CLI.

62 00:10:16.790 00:10:17.750 Pranav Narahari: So…

63 00:10:18.200 00:10:27.359 Pranav Narahari: like, let’s say Danny is trying to be… is trying to alias as Adam, they wouldn’t be able to use a GWS CLI for that portion of things.

64 00:10:27.560 00:10:30.940 Pranav Narahari: However, you could just use the API.

65 00:10:31.680 00:10:32.520 Pranav Narahari: So…

66 00:10:32.630 00:10:38.459 Pranav Narahari: Yeah, I remember that’s something you mentioned, like, on Friday, like, okay, wherever there’s gaps in CLI, let’s just use the API.

67 00:10:39.680 00:10:41.589 Uttam Kumaran: Yeah, so I just did this for us.

68 00:10:41.800 00:10:50.019 Uttam Kumaran: I have domain-wide delegation for a Brainforge platform application, For our account.

69 00:10:50.230 00:10:54.699 Uttam Kumaran: And I’m using the API to bring in all of those events in bulk.

70 00:10:55.230 00:11:05.939 Uttam Kumaran: Like, you can basically do the same thing, where you can create a Brainforge Google application, give it just the scopes that you need.

71 00:11:06.080 00:11:12.420 Uttam Kumaran: And then, on the fly, you can just pull using the API. It’ll be the exact same thing. I wouldn’t…

72 00:11:13.430 00:11:17.840 Uttam Kumaran: I wouldn’t, like, fuck around with, like, this PR or anything, like…

73 00:11:18.160 00:11:28.690 Uttam Kumaran: I don’t know. That’s whatever. Like, so… and… and to your point, you can actually just, like, have the call API and bring that in as just one-time context.

74 00:11:28.910 00:11:31.529 Uttam Kumaran: So, I can… if you want, I can…

75 00:11:32.480 00:11:42.939 Uttam Kumaran: I think I literally pushed the PR on Friday, which outlined, like, how I kind of did this, but the steps basically are… you just have to create, like,

76 00:11:44.140 00:11:50.359 Uttam Kumaran: I mean, you just have to create, like, a brain forge. I assume, like, for their service account, there’s a project in their Eden Google, right?

77 00:11:50.730 00:11:51.329 Pranav Narahari: Yep, yep.

78 00:11:51.330 00:11:56.970 Uttam Kumaran: Okay, so it’s the same thing, and you could just pull whatever you need through that. You’ll have, like, a…

79 00:11:57.080 00:12:08.759 Uttam Kumaran: the JSON file, that’s, like, the service account credentials, you can just use the API for that, and you can actually just apply domain-wide delegation to that, so…

80 00:12:09.320 00:12:12.019 Uttam Kumaran: Let me even see if I can…

81 00:12:13.300 00:12:16.160 Uttam Kumaran: Show you how we did ours, because it’s a…

82 00:12:16.160 00:12:16.950 Pranav Narahari: Same.

83 00:12:17.360 00:12:26.019 Pranav Narahari: part of what I think maybe is different, or maybe, like, you did this for Brainforge in a different way, is how would this work

84 00:12:26.840 00:12:35.770 Pranav Narahari: And what is the complexity to make this work for a chat interface? Because for a chat interface, let’s say how Danny’s gonna want to use this is, what is…

85 00:12:36.460 00:12:43.340 Pranav Narahari: the… Like, he could just say, what is going on for…

86 00:12:43.440 00:12:53.210 Pranav Narahari: this marketing effort. So, like, XYZ marketing effort, right? And so, how DWD would be best utilized for that is…

87 00:12:53.540 00:13:03.520 Pranav Narahari: And actually, let me back up first and, like, maybe get more information about DW, because you probably, since you used it for Brand Forge, you probably know more. Do you need to alias as a certain individual, or can you just bulk.

88 00:13:03.520 00:13:03.950 Uttam Kumaran: No.

89 00:13:03.950 00:13:05.750 Pranav Narahari: everybody’s information. So this is…

90 00:13:05.750 00:13:09.250 Uttam Kumaran: This is… this is the client ID. I’ve literally given it all the scopes.

91 00:13:09.650 00:13:22.559 Uttam Kumaran: Yeah. Which is, like, I wanted it to go through… have email access, a bunch of access, because I was working on spam, and I’m working on this email thing, and that’s all I did. So there’s this client ID, which is the Brainforge API client.

92 00:13:22.810 00:13:27.420 Uttam Kumaran: And I have that set up already under… under console.

93 00:13:27.740 00:13:29.909 Uttam Kumaran: So if you go to console.

94 00:13:30.850 00:13:33.459 Uttam Kumaran: You’ll see that there is a…

95 00:13:33.890 00:13:38.500 Uttam Kumaran: I have a BrainForge API, like, project,

96 00:13:43.110 00:13:43.959 Uttam Kumaran: I don’t know.

97 00:13:46.030 00:13:49.620 Uttam Kumaran: That’s our BrainForge platform.

98 00:13:50.790 00:13:54.259 Uttam Kumaran: This is it, right? Yep. This is the service account.

99 00:13:54.400 00:13:56.590 Uttam Kumaran: And this is… this is just by…

100 00:13:57.210 00:13:59.679 Uttam Kumaran: Client, there’s that client ID, right?

101 00:14:00.060 00:14:02.780 Uttam Kumaran: Yep. It’s 9-0, and that’s… Yep.

102 00:14:03.240 00:14:04.920 Uttam Kumaran: I’m pretty sure that’s, like.

103 00:14:07.490 00:14:24.329 Uttam Kumaran: One of these is this one. I don’t know which one it is, but I was actually… I actually just did a lot of this using the CLI, and so, yeah, so there’s… so there’s actually… but then this is where I would separate those two. Producing the report is different than allowing for the chat interface, right?

104 00:14:24.840 00:14:30.829 Uttam Kumaran: Like, the chat interface implies that you have static data sitting somewhere to call this over.

105 00:14:34.730 00:14:35.540 Pranav Narahari: Because…

106 00:14:35.540 00:14:39.600 Uttam Kumaran: Because otherwise, like, well, I mean, otherwise, you have to pull it

107 00:14:40.220 00:14:42.060 Uttam Kumaran: Yeah, I guess this is where I’m, like.

108 00:14:43.700 00:14:49.089 Uttam Kumaran: there’s… I feel like there’s 3 things going on. There’s one, like, there may be a need to look at, like.

109 00:14:49.320 00:14:51.279 Uttam Kumaran: Trent, like, just like I’m looking at

110 00:14:51.730 00:15:00.909 Uttam Kumaran: how many calendar events are over time, there may be a need for that. That’s not in scope, and there’s a need for, like, chatting over a set of data, right?

111 00:15:04.380 00:15:07.959 Uttam Kumaran: Or, the third thing is you can just chat over the report.

112 00:15:08.700 00:15:11.340 Uttam Kumaran: And the report can include, like.

113 00:15:11.970 00:15:15.239 Uttam Kumaran: enough information where you’re almost just, like, chatting with a PDF.

114 00:15:15.340 00:15:18.060 Uttam Kumaran: Right? Or chatting with, like, a long report.

115 00:15:18.330 00:15:19.260 Uttam Kumaran: But…

116 00:15:19.260 00:15:19.720 Pranav Narahari: Yeah.

117 00:15:19.720 00:15:25.680 Uttam Kumaran: Again, what’s gonna be the most flexible is, like, hey, there’s a…

118 00:15:26.340 00:15:32.129 Uttam Kumaran: And this is where, like, I guess I’m like, I don’t know why those guys, like, Awasht and Sam, didn’t just…

119 00:15:32.400 00:15:42.730 Uttam Kumaran: Why don’t they… I’m not understanding why they don’t have context here, because for me, Away should have proposed, like, hey, looks like you want to chat over a bunch of data, we could be landing this somewhere.

120 00:15:43.380 00:15:50.230 Uttam Kumaran: And then the theme-finding thing is an AI piece, right? For the AI to, like, be able to look through themes.

121 00:15:50.350 00:15:54.619 Uttam Kumaran: That does… that is, like, a… that’s, like, the Sam’s area for… for…

122 00:15:54.810 00:15:57.179 Uttam Kumaran: for looking through that, so this is where I’m, like.

123 00:15:57.300 00:16:00.189 Uttam Kumaran: My hunch is that you do have to land data somewhere now.

124 00:16:00.620 00:16:02.080 Uttam Kumaran: Right?

125 00:16:02.490 00:16:03.759 Uttam Kumaran: Chat over it.

126 00:16:04.040 00:16:12.380 Uttam Kumaran: even if it’s just a week’s worth, you have to put it somewhere. Wherever that is, we can debate that, but this isn’t, like, an on-the-fly thing, because…

127 00:16:14.760 00:16:15.510 Pranav Narahari: Mmm.

128 00:16:15.710 00:16:16.179 Uttam Kumaran: You close…

129 00:16:16.180 00:16:16.550 Pranav Narahari: Oh, yeah.

130 00:16:16.550 00:16:23.299 Uttam Kumaran: you could be… yeah, you could… I mean, yeah, I guess, like, every time then a chat starts, you want to pull that information in.

131 00:16:23.480 00:16:25.789 Uttam Kumaran: You might as well just store it at the start of the week.

132 00:16:27.500 00:16:29.790 Uttam Kumaran: For the past week, and then some… yeah.

133 00:16:29.790 00:16:30.570 Pranav Narahari: Yeah.

134 00:16:31.300 00:16:39.600 Pranav Narahari: I guess with, with, like, the Slack MCP and the GWO CLI, like.

135 00:16:40.150 00:16:41.400 Pranav Narahari: I think we could do it.

136 00:16:41.400 00:16:43.650 Uttam Kumaran: Oh, okay, yeah, fair.

137 00:16:43.650 00:16:47.810 Pranav Narahari: Right? Yeah. That’s where I’m like… I just, I don’t know, because I…

138 00:16:47.930 00:16:51.570 Uttam Kumaran: I just don’t know, how much data we’re talking about.

139 00:16:54.510 00:16:58.230 Uttam Kumaran: So this is where you have to do some testing, like, Yeah.

140 00:16:58.230 00:17:11.939 Pranav Narahari: Yeah, so this is what we’re doing today, right? And I think maybe I’ll have more context with that today, is like, I’m going to push Danny to, like… and he already kind of said that he would do this before he left, on his vacation, to give me…

141 00:17:12.060 00:17:16.059 Pranav Narahari: Like, what type of prompts are you looking to use the command center for?

142 00:17:16.060 00:17:17.030 Uttam Kumaran: Yeah.

143 00:17:17.369 00:17:21.540 Pranav Narahari: So, that’s gonna give me all the context, right? Because then I’m gonna be able to, from…

144 00:17:21.710 00:17:33.039 Pranav Narahari: from my side, from a more technical side, like, know, okay, what data do you need to be pulling in for this? And then, yeah, to your point, like, I’ll be able to, like, figure out what is the…

145 00:17:33.940 00:17:38.909 Pranav Narahari: how big of a data set are we trying to pull in? Like, what is going to be the context?

146 00:17:40.560 00:17:46.330 Uttam Kumaran: Yeah, I agree. Like, I think that… I think that’s gonna be important, because if he just wants to look at this past week’s.

147 00:17:46.500 00:17:51.840 Uttam Kumaran: Then, like, you can start the chat interface, you can run a script that basically, like.

148 00:17:52.350 00:17:56.189 Uttam Kumaran: Pulls it all into one place, just for that chat interface.

149 00:17:56.640 00:18:02.099 Uttam Kumaran: Yeah. But again, I think either way, like, if it’s more than a hundred… if it’s, like, 100,000 Slack messages.

150 00:18:02.460 00:18:05.140 Uttam Kumaran: Or, let’s say it’s, like, a thousand Slack messages.

151 00:18:06.880 00:18:12.430 Uttam Kumaran: you could… you could… there’s a middle ground. You could literally just store that in a… in a DuckDB,

152 00:18:12.740 00:18:14.589 Uttam Kumaran: Or, like, a SQLite table.

153 00:18:14.760 00:18:17.209 Uttam Kumaran: And then just, like, run your chat over that.

154 00:18:17.340 00:18:20.099 Uttam Kumaran: You don’t have to, like, store it long-term, right?

155 00:18:21.000 00:18:26.979 Uttam Kumaran: Because, again, what we’re debating is, like… because then you… because think about it, if you have a thousand messages, you have to then, what.

156 00:18:27.620 00:18:30.839 Uttam Kumaran: do, like, Slack MCP for a thousand times.

157 00:18:31.090 00:18:32.290 Uttam Kumaran: And then…

158 00:18:33.520 00:18:43.579 Uttam Kumaran: Yeah, like, I think that’s just what I want you to just avoid. That just doesn’t seem like a great solution. Like, for me, I’m running analytics on, like, hundreds and hundreds of events.

159 00:18:43.700 00:18:45.209 Uttam Kumaran: And I’m like, yeah, I’ll just…

160 00:18:45.380 00:18:47.660 Uttam Kumaran: It took me 2 seconds to just store all that.

161 00:18:48.040 00:18:52.839 Uttam Kumaran: Into Snowflake. But again, I’m running, like, long… I’m not doing anything about this week.

162 00:18:52.970 00:18:53.800 Uttam Kumaran: I’m, like, looking at.

163 00:18:53.800 00:18:54.130 Pranav Narahari: Yeah.

164 00:18:54.130 00:18:55.900 Uttam Kumaran: Oracle Analytics.

165 00:18:57.770 00:19:13.440 Pranav Narahari: Yeah, and also one thing that was interesting that you said is, like, yeah, we can have the Slack MCP, we can have the WCLI, but then we can also bake into the system that if you’re pulling in data from more than 7 days ago, that you just look at the reports that were generated.

166 00:19:13.930 00:19:14.320 Uttam Kumaran: That’s called.

167 00:19:14.320 00:19:30.330 Pranav Narahari: Not data… not raw data from Google, not raw data from Slack. If it’s more than 7 days old, then you just get whatever context you can from the reports. And that way, we’re not pulling in, yeah, you know, a ton of, you know, hundreds of thousands of.

168 00:19:30.540 00:19:38.780 Uttam Kumaran: But, like, kicking off, like, a small ETL job on Sunday to pull in all the data.

169 00:19:39.180 00:19:42.450 Uttam Kumaran: like, I could bill that for you right now, so…

170 00:19:42.570 00:19:44.340 Uttam Kumaran: I guess one thing I’m just, like.

171 00:19:44.800 00:19:52.539 Uttam Kumaran: I’m just getting a little bit annoyed that, like, Sam or Awash are not, like, helping you here, and, like, I don’t know what the fuck they’re doing, because…

172 00:19:52.660 00:19:54.340 Uttam Kumaran: This is, like, not in, like…

173 00:19:54.460 00:20:05.379 Uttam Kumaran: we’re, like, a month in, like, why aren’t they understanding, like, what this project is? So what I’m gonna do, I’m gonna send… I’m just gonna escalate, I’m gonna send a note, basically, like, hey guys.

174 00:20:05.540 00:20:12.690 Uttam Kumaran: there’s… there’s a couple of asks here. There’s, like, this report generation, and there’s this, like…

175 00:20:14.690 00:20:19.410 Uttam Kumaran: command center, where I need to ask questions over the data that went into the report.

176 00:20:20.010 00:20:20.400 Pranav Narahari: Yup.

177 00:20:20.400 00:20:25.719 Uttam Kumaran: Like, it’s been more than a month, and there’s been no solidification of a technical roadmap.

178 00:20:26.430 00:20:36.760 Uttam Kumaran: So, like, I’m gonna basically ask them to either solve that, or I’m just gonna move… they’re just gonna move off. Like, Sam’s just gonna not be… I don’t know, I’m not sure what I’m gonna do, because…

179 00:20:37.040 00:20:41.680 Uttam Kumaran: I think this is kind of ridiculous, like, that I have to sort of Think about this, but…

180 00:20:42.070 00:20:52.939 Uttam Kumaran: to tell you, I think… I think just assume that you can solve either of them. If what you need from me today, or tomorrow, is like, hey, can we run a test? I can do this for you.

181 00:20:53.350 00:20:55.580 Uttam Kumaran: Yeah.

182 00:20:55.760 00:21:10.099 Pranav Narahari: My… so why I messaged this morning is that I just want to be clear on the complexity of whatever decision… whatever path we go down, right? So there’s… there’s two paths that I see right now. We can stand up a DuckDB,

183 00:21:10.100 00:21:16.920 Pranav Narahari: data warehouse and have an ETL job to pull in the data on a weekly, daily basis, whatever.

184 00:21:17.000 00:21:20.869 Pranav Narahari: Or, potentially with the type of messages that

185 00:21:20.970 00:21:32.760 Pranav Narahari: Danny’s gonna be using the command center for. We’re not going to need to even stand that up, and we can just have GWS, CLI commands, and then Slack MCP do all of the work.

186 00:21:32.870 00:21:40.340 Pranav Narahari: And that’ll still be good enough for his purposes. Okay. If that’s the case, then it’s even simpler, right?

187 00:21:40.340 00:21:44.369 Uttam Kumaran: Why don’t you figure… why don’t you figure out the prompts that he’s trying to ask?

188 00:21:44.670 00:21:47.399 Uttam Kumaran: And then let’s try, we’ll try both strategies.

189 00:21:47.650 00:21:53.529 Uttam Kumaran: And I’ll show you what’s working. Do you have all… all of those things, like the… the…

190 00:21:53.750 00:21:57.860 Uttam Kumaran: the, service account… where is that? Is that all in one password?

191 00:21:58.230 00:22:08.249 Pranav Narahari: That’s, yes, it’s all in 1Password, it’s also in linear. There’s a ticket where I confirm that the service account is able to have domain-wide delegation.

192 00:22:08.250 00:22:11.719 Uttam Kumaran: But that JSON piece, is that in 1Password?

193 00:22:12.910 00:22:14.459 Pranav Narahari: The client ID, you’re talking about?

194 00:22:14.460 00:22:16.769 Uttam Kumaran: Yeah, well, you get that client JSON, right?

195 00:22:17.540 00:22:21.559 Pranav Narahari: Yeah, yeah, yeah, it is, and let me.

196 00:22:21.560 00:22:31.310 Uttam Kumaran: Can you just check? Because I have a feeling that I may need to help you here if, like, they’re not gonna get back to you, so, like, I just want to be able to, like, later tonight or something, like, test a couple things out.

197 00:22:31.450 00:22:39.219 Uttam Kumaran: If you, like, for example, if you’re today were to be, like, giving me these prompts, like, here are some of the prompts, I may be able to just run some tests on my own.

198 00:22:40.000 00:22:40.700 Pranav Narahari: Yeah.

199 00:22:58.580 00:23:02.939 Pranav Narahari: Yes, it’s gonna be in our GCP service, for sure.

200 00:23:20.910 00:23:23.329 Uttam Kumaran: And this is sort of like a weekly theme report.

201 00:23:25.320 00:23:25.870 Pranav Narahari: Yep.

202 00:23:26.220 00:23:27.090 Uttam Kumaran: Okay, perfect.

203 00:23:27.720 00:23:28.280 Pranav Narahari: Yeah.

204 00:23:32.900 00:23:35.959 Pranav Narahari: But also, my understanding is that it…

205 00:23:36.110 00:23:39.309 Pranav Narahari: What is the additional complexity if we make it a daily report?

206 00:23:40.800 00:23:42.300 Uttam Kumaran: No, I don’t think there is any.

207 00:23:42.300 00:23:46.900 Pranav Narahari: There’s nothing right, yeah, yeah. So my… I think the thing is, is just, like, it doesn’t need to be…

208 00:23:47.060 00:23:59.490 Pranav Narahari: generated on the fly within a chat interface. All it… because, so, what that tells me, or then what… how that reduces, you know, the technical complexity, or just…

209 00:23:59.610 00:24:04.040 Pranav Narahari: The… makes a simpler product is that this can be done…

210 00:24:04.160 00:24:11.670 Pranav Narahari: this could take… this could take, you know, 3 minutes, 5 minutes, an hour to actually generate, if needed.

211 00:24:11.810 00:24:15.560 Pranav Narahari: You know, if it’s in a chat interface, it asks, do we have to worry about execution time and stuff.

212 00:24:16.320 00:24:21.870 Uttam Kumaran: Yeah, I agree. So that’s what I’m saying, if you could just deliver me, like, what the prompts are.

213 00:24:21.970 00:24:25.619 Uttam Kumaran: then I’ll try to build with you as some type of demo.

214 00:24:25.970 00:24:30.460 Uttam Kumaran: That shows it working for those prompts, and then we can pick the technical avenue.

215 00:24:31.160 00:24:32.090 Pranav Narahari: Perfect.

216 00:24:32.310 00:24:37.640 Pranav Narahari: Yeah, I mean, and also… Danny’s been out for 2 weeks, right? So basically, as soon as the

217 00:24:37.670 00:24:55.490 Pranav Narahari: project started, he was out. And so today is the first time I’m actually going to be able to have a conversation with him. Hopefully, like, fri- so we had a whole conversation on Friday, right, to, like, prepare me for that meeting. He ended up having to cancel it, because, he’s just getting back into office, like, he has other stuff.

218 00:24:55.620 00:24:57.100 Pranav Narahari: Maybe it’s a good sign.

219 00:24:57.100 00:24:59.920 Uttam Kumaran: So do you have stuff to, like, kind of share with him today?

220 00:25:00.550 00:25:05.839 Pranav Narahari: Yeah, yeah, I have a ton of stuff, because I’m guessing he’s not caught up on Slack, so I have to even get his…

221 00:25:05.980 00:25:16.560 Pranav Narahari: his whole setup, going. Okay, okay. Yeah, I have a ton of stuff to talk about today. Also, I want to talk about, like, the new chat interface that, like, mock-up that I built, like, that lovable.

222 00:25:16.700 00:25:23.089 Pranav Narahari: So, I do feel pretty confident, but I just… I’m going to also talk about…

223 00:25:23.200 00:25:31.480 Pranav Narahari: the service account piece of things, right? And one concern that Sam kind of had after I was talking about it was that we’re not going to be able to…

224 00:25:31.700 00:25:35.570 Pranav Narahari: use the service account in the way that we thought we could.

225 00:25:35.570 00:25:40.099 Uttam Kumaran: No, you said, that’s not true at all, yeah. This is where I’m just like…

226 00:25:41.740 00:25:50.620 Uttam Kumaran: I’m gonna solve… I’m just gonna… I’m gonna get involved this week and solve these things, because, like, I don’t want you to be left out of dry, but this is, like, kind of disappointing.

227 00:25:50.910 00:25:53.580 Uttam Kumaran: That, like, I’m not… I’m having to, like, do this.

228 00:25:53.770 00:25:56.909 Uttam Kumaran: But I’m gonna send a note right now, basically saying, like.

229 00:25:57.650 00:26:00.330 Uttam Kumaran: it’s so obvious you can use the Google

230 00:26:00.770 00:26:10.089 Uttam Kumaran: admin API to just pull. The ultimate decision here is, like, do we need to backfill and pull all this shit, or can you just pull a week’s worth?

231 00:26:10.270 00:26:17.300 Uttam Kumaran: Chove it into a database, produce the report, throw both of those into context for the Command Center.

232 00:26:19.190 00:26:21.640 Uttam Kumaran: How… it’s not any more difficult than that.

233 00:26:22.460 00:26:29.770 Uttam Kumaran: maybe I’m just, like, Yeah, so let me send a note about that. My… that’s my proposal, is, like.

234 00:26:31.090 00:26:38.569 Uttam Kumaran: like, pull whatever data is needed to run the report. The report and that data sits just in context behind for the chat.

235 00:26:39.010 00:26:42.839 Uttam Kumaran: And once I get your prompts.

236 00:26:42.970 00:26:44.879 Uttam Kumaran: It’ll be easy for me to validate that.

237 00:26:47.600 00:26:48.180 Pranav Narahari: Okay.

238 00:26:55.820 00:26:56.700 Pranav Narahari: Yay.

239 00:26:58.190 00:27:00.759 Pranav Narahari: I’m looking for the client ID…

240 00:27:03.150 00:27:07.009 Pranav Narahari: It’s like, usually the JSON comes, like, client underscore secret, right?

241 00:27:08.200 00:27:10.940 Uttam Kumaran: It should be, like, a file.

242 00:27:11.360 00:27:13.520 Pranav Narahari: Yeah, client underscore secret dot JSON.

243 00:27:14.580 00:27:16.239 Uttam Kumaran: Yeah, it should be somewhere like that.

244 00:27:16.240 00:27:19.789 Pranav Narahari: Okay, so then it is in one pass, then, I’m pretty sure. Okay.

245 00:27:20.540 00:27:23.529 Pranav Narahari: Yes, it’s, yeah, I also need a direct link.

246 00:27:23.530 00:27:25.060 Uttam Kumaran: And there’s a repo for this?

247 00:27:26.010 00:27:26.610 Pranav Narahari: Yeah.

248 00:27:27.370 00:27:32.919 Uttam Kumaran: Okay, can you make sure… yeah, if you can send me that, or make sure I’m there, that way, in case I trust anything, I’ll just put it in there.

249 00:27:34.760 00:27:37.249 Pranav Narahari: Oh, wait, repo as in, like, a 1Pass?

250 00:27:38.250 00:27:39.630 Uttam Kumaran: Yes. Or, like…

251 00:27:39.630 00:27:42.369 Pranav Narahari: Oh, yeah, yeah, it’s just the Eden one. We’re all just…

252 00:27:42.370 00:27:46.230 Uttam Kumaran: No, no, no, I mean, actual GitHub repo, yeah.

253 00:27:46.230 00:27:48.889 Pranav Narahari: Yeah, yeah, we have one as well, I’ll link that as well.

254 00:27:51.800 00:27:54.420 Pranav Narahari: Here’s the one pass for the client secret.

255 00:27:57.950 00:28:01.100 Pranav Narahari: Yeah, I mean, I guess, you know, I also want to take…

256 00:28:01.270 00:28:08.250 Pranav Narahari: and this is kind of the same feedback that you kind of gave me last week, which I’m gonna get better at, and I tried to do that today, too, is like.

257 00:28:08.560 00:28:12.750 Pranav Narahari: I need to do a better job of knowing when, like, the SL is not…

258 00:28:12.930 00:28:15.360 Pranav Narahari: Doing what they should be doing, you know?

259 00:28:15.790 00:28:16.640 Pranav Narahari: like…

260 00:28:17.240 00:28:22.079 Uttam Kumaran: Yeah, like, I don’t know why, this is, like, something I just cleared up in, like, 15 minutes.

261 00:28:22.320 00:28:25.250 Uttam Kumaran: And I have, like, the least understanding about this project.

262 00:28:25.460 00:28:30.850 Uttam Kumaran: I’m gonna call people out right now, I’m gonna basically, basically tell Sam… I’m basically also gonna tell Awash that, like.

263 00:28:31.390 00:28:36.370 Uttam Kumaran: there is… Data engineering and data movement.

264 00:28:36.980 00:28:38.560 Uttam Kumaran: on this project.

265 00:28:39.280 00:28:45.980 Uttam Kumaran: Okay, so let me send the… let me send the message, get the prompts for me, and then I’ll help you get something early this week, so…

266 00:28:46.190 00:28:50.209 Pranav Narahari: Okay, cool. Yeah, I mean, I’ll, I have a meeting at 2.30.

267 00:28:50.400 00:28:57.089 Pranav Narahari: with, Danny. Hopefully he makes it. He still hasn’t accepted it. I sent a message in the external chat, he hasn’t responded to that yet either, so…

268 00:28:57.350 00:29:02.290 Pranav Narahari: I think it could get moved to tomorrow. I asked Adam, too, who’s, like.

269 00:29:02.900 00:29:08.319 Pranav Narahari: he’s, you know, he’s the IT lead who’s been a lot more responsive with me, but he’s also kind of like…

270 00:29:08.510 00:29:22.359 Pranav Narahari: he’s lacking, kind of, understanding as well of how Danny would want to use this command center, which is not a good… which is not great. Because I want him, Adam, to be, like, the one to fully understand, like, where Danny wants to use this project.

271 00:29:25.330 00:29:30.660 Pranav Narahari: Because he’s gonna be way more responsive to me. And he’s like, you know, like, how you mentioned, like, how there’s, like, the…

272 00:29:30.770 00:29:36.569 Pranav Narahari: the main stakeholder on the client side, but then also the champion. Like, Adam is supposed to be the champion, so I want…

273 00:29:36.730 00:29:44.800 Pranav Narahari: I don’t want, kind of, him to be like, hey, how do we use the command center? He should be the one, like, trying to sell the command center internally to Eden, right?

274 00:29:44.800 00:29:45.380 Uttam Kumaran: Yeah.

275 00:29:45.870 00:29:46.460 Pranav Narahari: Yeah.

276 00:29:46.600 00:29:52.870 Pranav Narahari: So, that prompts thing should clear up that a lot, because Adam does like this command center,

277 00:29:53.220 00:29:56.920 Pranav Narahari: He is very responsive, he is, like…

278 00:29:57.400 00:30:05.579 Pranav Narahari: He is kind of… he’s on board, he likes the idea, but in terms of the actual usage of it, he’s lacking.

279 00:30:05.930 00:30:07.869 Pranav Narahari: It’s lacking a little bit of clarity on that.

280 00:30:08.030 00:30:09.060 Uttam Kumaran: Okay, events.

281 00:30:09.060 00:30:11.760 Pranav Narahari: Yeah, so… just waiting on Danny, basically.

282 00:30:12.840 00:30:13.460 Uttam Kumaran: Okay.

283 00:30:13.980 00:30:14.510 Pranav Narahari: Yeah.

284 00:30:16.530 00:30:19.280 Pranav Narahari: Okay, I got… I sent over both those things.

285 00:30:19.910 00:30:23.660 Uttam Kumaran: Okay. Okay, let me send this message, and then, yeah, I’ll talk to you then later today.

286 00:30:24.190 00:30:25.299 Pranav Narahari: Okay, yup, yeah.

287 00:30:25.300 00:30:26.740 Uttam Kumaran: Okay. Thanks, dude. Bye.

288 00:30:27.210 00:30:27.960 Pranav Narahari: Yeah, bye.