Meeting Title: Brainforge x Eden Remo DevOps Sync Date: 2026-05-05 Meeting participants: Awaish Kumar, Robert Tseng
WEBVTT
1 00:00:52.860 ⇒ 00:00:53.710 Robert Tseng: 8.
2 00:00:54.270 ⇒ 00:00:54.800 Awaish Kumar: Hello.
3 00:00:55.360 ⇒ 00:01:02.460 Robert Tseng: Okay, yeah, so… Man, what a… what a day.
4 00:01:02.660 ⇒ 00:01:15.930 Robert Tseng: Yeah, I mean, I… I think I was trying to, like, do my best at, like, understanding, like, what your take was in trying to put together this… this doc. So, I mean, you know, we’re basically trying to define, like, how they should be…
5 00:01:16.520 ⇒ 00:01:27.469 Robert Tseng: relying on us for deployment. So, I was thinking, okay, they have this, like, random repo that they have. It’s not as much as Adam thinks it is, he’s just kind of, like, riffing.
6 00:01:28.910 ⇒ 00:01:37.470 Robert Tseng: Yeah, I mean, a lot of… a lot of it isn’t actually production. Like, it’s not really code. It’s just, like, a bunch of ideas, and
7 00:01:38.360 ⇒ 00:01:55.490 Robert Tseng: yeah, if we’re going to kind of adopt your approach of infrastructure as code, really kind of orchestrate everything through Dagster, then I took the spirit of that and tried to create this delivery model, where I’m like, look, Adam, you can have your own repo where you just ship whatever the heck you want.
8 00:01:55.570 ⇒ 00:02:08.199 Robert Tseng: doesn’t matter, because we still have our managed production environment. We have a couple repos. We have mostly the analytics, and then also there’s, like, Eden Remo. I mean, I don’t know which ones we actually maintain, but I’m just trying to.
9 00:02:08.520 ⇒ 00:02:09.139 Awaish Kumar: Okay.
10 00:02:09.810 ⇒ 00:02:10.490 Robert Tseng: Say that.
11 00:02:10.860 ⇒ 00:02:18.489 Robert Tseng: And then, as you read through it, it’s like, okay, so they can keep shipping stuff to us,
12 00:02:18.800 ⇒ 00:02:32.959 Robert Tseng: And for us to… but we need to have, like, a more programmatic way of, like, sifting through, you know, their changes. Because he’s right, he shipped 1,000 commits. There’s no way that me and Zoran are going to review 1,000 commits for him.
13 00:02:32.960 ⇒ 00:02:42.839 Robert Tseng: And so, you know, most of those commits are not really things that impact our, the platform that we’re maintaining, but we should be, like, kind of…
14 00:02:42.970 ⇒ 00:02:49.879 Robert Tseng: kind of setting up a review process, kind of similar to how we allow people from all across Brainforge to be able to ship
15 00:02:49.880 ⇒ 00:03:09.719 Robert Tseng: changes into Brainforce Platform. You know, it runs through all of these, kind of, like, these checks, and there is, like, not every PR is, like, manually reviewed, right? So, there may be, like, a DevOps piece to this in order to… and then, obviously, the orchestration side, like, you want to… you want to coordinate it all through Dexter as opposed to GitHub Actions, or whatever we’re using currently. So,
16 00:03:09.820 ⇒ 00:03:15.050 Robert Tseng: Yeah, that’s kind of what I’ve tried to articulate here. There’s, like, kind of the,
17 00:03:15.520 ⇒ 00:03:24.290 Robert Tseng: Is their… their repo, which is whatever the heck this architecture implementation thing is, and then our primary repos, which…
18 00:03:24.420 ⇒ 00:03:38.379 Robert Tseng: I don’t think this is right, because we don’t really maintain this, but this is where I need you to kind of, like, finish out this doc, and, like, what are we committing to? So, there’s, like, a four-week rollout where it’s like, okay, over this month, since we’re month-to-month right now.
19 00:03:38.390 ⇒ 00:03:50.510 Robert Tseng: we need to really define clearly what gets passed into, production. So, there’s, like, yeah, I mean, all this is, like, random AI jargon here in terms of, like, what these terms are, but
20 00:03:50.760 ⇒ 00:04:06.490 Robert Tseng: hopefully you can try to see, that I’m trying to basically get to some sort of, like, review cadence, where I’m able to, you know, when they’re asking, when they’re just shipping us, like, a random, like, hey, look at this random message, you know, like.
21 00:04:06.490 ⇒ 00:04:11.029 Robert Tseng: all this friend crap that I get from… from… from them.
22 00:04:11.750 ⇒ 00:04:18.009 Robert Tseng: like, random plan, can you take this on? Random, just, like, dump of rap. And there were, like, this…
23 00:04:18.149 ⇒ 00:04:36.820 Robert Tseng: shares this random link. Like, I can’t just, like, go stop what I’m doing and go review that. It needs to get pushed through our, process, and things that do actually need my review will kind of be, like, I’m assuming, like, it… I’ll get, like, something actually reviewable.
24 00:04:36.860 ⇒ 00:04:43.129 Robert Tseng: Rather than wasting my time trying to go and basically running this GitHub through agents and trying to
25 00:04:43.130 ⇒ 00:04:47.209 Robert Tseng: asked, like, what was the most recent change, what are the themes? I was like.
26 00:04:47.210 ⇒ 00:05:03.859 Robert Tseng: I’m, like, having to start from ground zero every time. Like, I get a random kind of ping like this. So, I mean, anyway, I think you can kind of understand, like, what I’m trying to do here. What questions do you have? Like, well, that’s… that’s what I really want to kind of nail down on this call.
27 00:05:03.860 ⇒ 00:05:04.530 Awaish Kumar: Nope.
28 00:05:04.530 ⇒ 00:05:05.080 Robert Tseng: Yeah.
29 00:05:05.600 ⇒ 00:05:09.800 Awaish Kumar: So, number one thing is that, like, what, we are trying to achieve
30 00:05:10.000 ⇒ 00:05:13.169 Awaish Kumar: Regarding his… I think he seems…
31 00:05:13.350 ⇒ 00:05:22.790 Awaish Kumar: number one, from my observation, it seems like to have speed with agents. Like, use the eye, push things out, a lot of commutes and… and things like that.
32 00:05:24.450 ⇒ 00:05:30.049 Awaish Kumar: Second thing regarding the… what I… what I’m thinking about is more infrastructure.
33 00:05:30.960 ⇒ 00:05:39.249 Awaish Kumar: one piece. But second thing is that when we are saying, like, what he’s, like, throwing, this, like, text to us, like, it’s like,
34 00:05:39.420 ⇒ 00:05:47.390 Awaish Kumar: Like, they can always do it. We don’t have any way to, like, handle that, right? Even if they push in a PR, like, what we are going to do with our text.
35 00:05:47.590 ⇒ 00:05:48.990 Awaish Kumar: It’s just text, right?
36 00:05:49.680 ⇒ 00:05:52.129 Awaish Kumar: And there’s… Right.
37 00:05:52.130 ⇒ 00:06:07.419 Robert Tseng: You’re right, it’s just text, so it doesn’t… there’s no code to it. He’s just, like, pushing… he has this giant repo of, like, ideas. That’s what I think it is. And, like, none of it is really hooked up to anything that we maintain, like, it’s so… Yeah, it’s just, like, it’s just… it’s just text, you’re right.
38 00:06:08.100 ⇒ 00:06:18.620 Awaish Kumar: And then, like, I just went through it, for example, the one… for this example, and created a plan that I can… that can actually be executed with the tickets. So…
39 00:06:18.790 ⇒ 00:06:21.889 Awaish Kumar: Yeah, coming back to this document,
40 00:06:22.650 ⇒ 00:06:33.259 Awaish Kumar: like, I get it, but the only thing I’m concerned about is scope, right? When we are… when I read all these things, it does not include any kind of,
41 00:06:33.860 ⇒ 00:06:34.980 Awaish Kumar: imp… like…
42 00:06:35.120 ⇒ 00:06:46.289 Awaish Kumar: implementation or shifting to the IAC. It’s more like defining role-based access for, for example, platforms you’re using, defining checks on GitHub.
43 00:06:46.450 ⇒ 00:06:47.900 Awaish Kumar: And all those things.
44 00:06:48.840 ⇒ 00:06:57.250 Awaish Kumar: Like, how… PR gets, approved, and guidelines for that, and all those sorts of things, right?
45 00:06:58.860 ⇒ 00:07:03.939 Robert Tseng: Okay, so your feedback is this is more DevOps, doesn’t actually include the infrastructure risk code?
46 00:07:05.810 ⇒ 00:07:07.980 Awaish Kumar: Yes, like, it does not include how…
47 00:07:08.430 ⇒ 00:07:15.569 Awaish Kumar: for a segment, like, there are two pieces, right? Number one is he… he has defined a config that is…
48 00:07:15.780 ⇒ 00:07:21.430 Awaish Kumar: just his idea, like, AI-generated. It’s not a standard of anything, right?
49 00:07:21.580 ⇒ 00:07:22.420 Robert Tseng: Yeah.
50 00:07:22.500 ⇒ 00:07:34.329 Awaish Kumar: there is, like, we need… then we need to first agree on that, like, what are the building blocks of a config? Because AI can generate one type, a different config, another time, and a different one.
51 00:07:34.360 ⇒ 00:07:45.599 Awaish Kumar: So, like, that’s kind of defining all the config, what building blocks will be going into that config, how it gets mapped to the segment,
52 00:07:45.780 ⇒ 00:07:49.350 Awaish Kumar: Source, what’s his definition, what’s his sinks, and all those things.
53 00:07:49.600 ⇒ 00:07:55.799 Awaish Kumar: Right? And then writing this down all in a code. So that’s basically moves…
54 00:07:55.950 ⇒ 00:08:03.729 Awaish Kumar: the manual UI changes into the code. I don’t see, like, any… that kind of changes in this document.
55 00:08:04.270 ⇒ 00:08:21.789 Robert Tseng: Okay, yeah, I mean, can you go ahead and edit it? Like, this is my understanding of how it works, so… I… I think this is the best I can do. Like, I already spent, like, almost 45 minutes on this, so, like, you can strike out anything that you don’t feel like is important. Like, I… I just…
56 00:08:22.130 ⇒ 00:08:26.850 Robert Tseng: If you think I was too heavy-handed on defining the DevOps steps, you can remove that.
57 00:08:26.870 ⇒ 00:08:32.500 Robert Tseng: you know, these are really for us. Like, I just want them to have a clear SOP for, like.
58 00:08:32.500 ⇒ 00:08:48.970 Robert Tseng: when I ship random AI shit to Brainforge, like, this is what’s gonna happen. Like, they’re going to need to set, you know, these things… they need to set up basic, like, standard configs, you know, in order to do… move to infrastructure as code. I agree, like, we have to have some basic building blocks and, like, kind of the standards around that.
59 00:08:48.970 ⇒ 00:09:03.699 Robert Tseng: So maybe that’s actually what week one should be. Maybe week two is more, okay, then we do the roles access permissions. You know, this stuff is really just DevOps stuff, like, do we really need two weeks of DevOps stuff? Like, probably not. Like, I think there’s, like, some…
60 00:09:03.700 ⇒ 00:09:05.009 Awaish Kumar: Some, revisions.
61 00:09:05.010 ⇒ 00:09:06.250 Robert Tseng: here? Yeah.
62 00:09:06.890 ⇒ 00:09:11.359 Awaish Kumar: One is… like, GitHub and PR, that is in…
63 00:09:11.760 ⇒ 00:09:15.699 Awaish Kumar: why I’m saying infrastructure as a code is that we are not just doing,
64 00:09:16.020 ⇒ 00:09:19.740 Awaish Kumar: infries a code, really, like, it’s… it’s not just, like,
65 00:09:20.110 ⇒ 00:09:28.939 Awaish Kumar: setting up class… roles and things, right? We are completely rewriting the sec… like, all the source… segment sources, connections via SDK.
66 00:09:29.130 ⇒ 00:09:29.730 Awaish Kumar: will be.
67 00:09:29.730 ⇒ 00:09:30.500 Robert Tseng: Yes.
68 00:09:30.500 ⇒ 00:09:36.880 Awaish Kumar: creating a mapping between user-defined configs to actually the REST API requests that will.
69 00:09:36.880 ⇒ 00:09:54.420 Robert Tseng: Yeah, they want to basically kill the UI. They want to move completely to code. Yeah, I agree. And they, like, may not stay on segment, right? If that’s the case, then we would just delete segment probably by July, if we’re still its client. We would maybe move to router stack, or something that’s just, like, pure backend. So, yeah, I think I understand.
70 00:09:54.420 ⇒ 00:10:03.679 Robert Tseng: why that’s important, I don’t really know why that’s important for them. It’s not like they’re really doing that much in segment. But yeah, I think… I think they just…
71 00:10:03.970 ⇒ 00:10:10.030 Robert Tseng: There’s this, like, addiction that they have to speed right now, and they just want to push a bunch of stuff.
72 00:10:10.240 ⇒ 00:10:20.649 Robert Tseng: we… but we’re gonna get caught with, like, we just have to keep reviewing their stuff. So I… you know, I… I don’t… I think there’s a balance here of, like,
73 00:10:21.520 ⇒ 00:10:39.530 Robert Tseng: Yeah, being pushed back where you feel like it’s unnecessary, but, like, you can see where they’re headed. They are moving towards a future where they just want to deploy agents on everything, and they just want to keep us just for governance. So, like, that’s… I think that’s the future. Like, that’s… that’s, like… they don’t… there’s no world where we are, like, having…
74 00:10:39.530 ⇒ 00:10:53.179 Robert Tseng: one owner for segment, one owner for GTM, one owner for the data warehouse. They’re trying to… they’re, like, taking a bulldozer and, like, smashing down all of those walls. Which is fine, because I think for us, if we set this up properly, the way I think about it is…
75 00:10:53.180 ⇒ 00:11:02.319 Robert Tseng: Maybe this is just the package that we deploy out of the box for future clients. If we have to run into a situation where we’re being asked to manage marketing data on a CPG brand or whatever.
76 00:11:02.320 ⇒ 00:11:06.489 Robert Tseng: Yeah, then I can only… then I only need, like, you know.
77 00:11:06.540 ⇒ 00:11:14.939 Robert Tseng: you come in, you set it up, then maybe Zoran can run this by himself. Zoran plus, like, one other person. So, you know, I think I’m trying to, like.
78 00:11:15.320 ⇒ 00:11:31.879 Robert Tseng: I do agree that, like, we should move to a more composable kind of, like, approach, where we’re not, like, having to designate owners for different parts of the stack. Like, we have to just knock it all down and have, like, a way to orchestrate, like, the…
79 00:11:31.880 ⇒ 00:11:46.180 Robert Tseng: the changes that the client will be pushing alongside us. So, I think however we get there, I think this becomes, like, our data governance as a service offering. So I’m just, like, kind of…
80 00:11:46.180 ⇒ 00:11:58.610 Robert Tseng: you know, talking a lot about this with you, because you being the SL and on the data side, like, yeah, like, I know this isn’t, like, your typical Snowflake or BigQuery, like, managed service setup, but, like, I’m trying to, like.
81 00:11:58.800 ⇒ 00:12:17.289 Robert Tseng: push the boundary of, like, is this even feasible? Can we run an engagement that’s purely just doing… doing this? And if you’re like, you know, this is too much, like, I don’t… I don’t think we should do that, or then, you know, let me know. But, like, I want to be able to articulate, like, what does that look like, and then we can decide
82 00:12:17.290 ⇒ 00:12:27.640 Robert Tseng: you know, if we want to take this across to other clients, too. Like, that’s, like, the business perspective for me, why I’m spending time, like, trying to figure this out.
83 00:12:29.320 ⇒ 00:12:30.130 Awaish Kumar: Okay.
84 00:12:30.420 ⇒ 00:12:40.429 Awaish Kumar: Okay, yeah, like, obviously you’re thinking about business things, and I’m thinking about, like, how we actually can make it… Yeah. …on the technical side.
85 00:12:40.590 ⇒ 00:12:46.340 Awaish Kumar: So, my things are, like, number one, that now you’re… now we are, like.
86 00:12:46.590 ⇒ 00:12:54.399 Awaish Kumar: thinking about more market thing. So, is this… what is the scope of this? Like, if I have to migrate everything from segment?
87 00:12:54.540 ⇒ 00:12:56.010 Awaish Kumar: into IEAC.
88 00:12:56.920 ⇒ 00:12:57.490 Awaish Kumar: in my.
89 00:12:57.490 ⇒ 00:12:58.150 Robert Tseng: Yeah.
90 00:12:58.340 ⇒ 00:13:01.180 Awaish Kumar: I don’t know, it’s my going to take longer, because
91 00:13:01.700 ⇒ 00:13:09.629 Awaish Kumar: like, Basque webhooks, there’s a lot of things there, like mixed panel connections, CIO connections, and…
92 00:13:10.770 ⇒ 00:13:11.579 Awaish Kumar: Lots of different…
93 00:13:11.580 ⇒ 00:13:17.210 Robert Tseng: Also, here’s the thing, like, what if you don’t do it? Like, what if you just, like, let them do it?
94 00:13:19.990 ⇒ 00:13:20.710 Awaish Kumar: Okay.
95 00:13:21.620 ⇒ 00:13:33.209 Robert Tseng: You know what I mean? Like, they have their plan, let their Asians go rip it, they can go and break stuff, I don’t really care, like, we’ll still, like, maintain, like, what we have right now, but, like, if they want to go and…
96 00:13:34.240 ⇒ 00:13:37.770 Robert Tseng: But, like, we… can we, like, build something that just, like…
97 00:13:37.890 ⇒ 00:13:50.460 Robert Tseng: Yeah. I mean, doesn’t necessarily tell them what to do, but, like, tells them what would be a breaking change, like, kind of filters through, like, you know, just let them kind of build that thing in parallel. Like, I don’t think we have to, like, you know.
98 00:13:50.460 ⇒ 00:13:57.130 Awaish Kumar: Yeah, for that, I… I would suggest, like, for example, we don’t target the existing,
99 00:13:57.500 ⇒ 00:14:07.059 Awaish Kumar: like, direct… bringing in everything from segment. We target, like, creating new sources, like, for example, for Martech, only for Martech, we might have just one or two
100 00:14:07.570 ⇒ 00:14:14.499 Awaish Kumar: sources that we can do. Like, that will help us come up with a template where you can actually
101 00:14:14.630 ⇒ 00:14:17.980 Awaish Kumar: Create corrections for segment, define configs.
102 00:14:18.120 ⇒ 00:14:21.670 Awaish Kumar: run dbt on top of it, and finally,
103 00:14:21.950 ⇒ 00:14:34.709 Awaish Kumar: deploy it, and run it, GitHub Actions or whatever. So that will be a complete package, and after that, we can define our PR checks and rules and everything. So that way,
104 00:14:34.850 ⇒ 00:14:39.209 Awaish Kumar: Then they can start adding new sources, connections, whatever they want to bring in.
105 00:14:40.040 ⇒ 00:14:47.269 Robert Tseng: Okay, that is… that makes sense. Then, where do you recommend we start? Which, which, which, which part should we… should we… should we run that way?
106 00:14:48.630 ⇒ 00:14:53.170 Awaish Kumar: I think we just define the scope in a way that we say, like, we are not,
107 00:14:53.510 ⇒ 00:14:58.140 Awaish Kumar: Migrating all the connections from…
108 00:14:58.280 ⇒ 00:15:03.310 Awaish Kumar: like, Scope is just Martech, so we are going to just bring in the cloud…
109 00:15:03.990 ⇒ 00:15:07.529 Awaish Kumar: the edge layer data that is coming in from MarTech stuff.
110 00:15:07.690 ⇒ 00:15:19.550 Awaish Kumar: Those will be our sources, whatever the config… we will define the config with them. We will define these sources, whatever one, two sources are there, maybe one or two destinations are there.
111 00:15:19.690 ⇒ 00:15:25.179 Awaish Kumar: And, then build on top of it. DVD models related to marketing, and then…
112 00:15:25.790 ⇒ 00:15:29.839 Awaish Kumar: the orchestration using CICD Gitavache.
113 00:15:30.570 ⇒ 00:15:46.199 Robert Tseng: Yeah, I think that’s fair, and I actually think that MarTech makes sense, because that data is already so dynamic and it’s broken all the time, I think it’s less risky. Like, I think all the stuff that we have maintained on the operations side, transactions, all the best bet hooks.
114 00:15:46.230 ⇒ 00:15:57.870 Robert Tseng: understood that we have to keep that up, because obviously we want to make sure that that’s… that’s reliable data. But for MarTech, it seems like speed, like, matters a lot to them, so if we can just define, like.
115 00:15:57.940 ⇒ 00:16:08.129 Robert Tseng: Okay, this was limited to just Martech, which is kind of why I only called it Martech. So maybe you have to add some, like, language here to limit that.
116 00:16:08.390 ⇒ 00:16:22.150 Robert Tseng: Yeah, maybe you help, kind of, write that out. Yeah, we’re just limiting this to only Martech. So, like, if you don’t know where to add one, at least just, like, give me some of the verbiage, and I can try to go in and edit it later, but, like.
117 00:16:22.220 ⇒ 00:16:31.960 Robert Tseng: Yeah, I think that’s… that’s… I’m… I’m… I agree with you that this should only be isolated to Martech. So, but to me, Martech includes…
118 00:16:32.150 ⇒ 00:16:43.949 Robert Tseng: you know, all visitors… yeah, I mean, edge layer… sure, edge layer tracking visitor behavior, but it’s also going to be the B… be, like, the product analytics work that Greg is kind of deploying as well with the team, which is, like.
119 00:16:43.950 ⇒ 00:17:04.820 Robert Tseng: you know, the actions that are being taken on their site, you know, intake, intake drop-off, stuff like that. So, anything that’s engaging with the marketing site before a transaction is placed. Everything past the purchase, you know, this is still gonna go through, like, our existing system. We’re not gonna migrate everything, but can we just say that everything, you know, MarTech-related can be put into this
120 00:17:04.819 ⇒ 00:17:05.710 Robert Tseng: approach.
121 00:17:06.540 ⇒ 00:17:12.750 Awaish Kumar: Yeah, so, like, what we are going to say is, like, we need dbt inside the same project.
122 00:17:12.970 ⇒ 00:17:14.200 Awaish Kumar: Right, Mark?
123 00:17:14.380 ⇒ 00:17:15.979 Awaish Kumar: So when we go from
124 00:17:16.430 ⇒ 00:17:22.559 Awaish Kumar: ingesting the data for marketing. We combine it with the orders data. So we basically need those
125 00:17:23.920 ⇒ 00:17:37.249 Awaish Kumar: like, models in there, right? So, like, we should have just one dbt project. So inside of our dbt project, we can include this, or we can move our dbt project into this one, but that will include all the
126 00:17:37.370 ⇒ 00:17:39.380 Awaish Kumar: other data models.
127 00:17:39.720 ⇒ 00:17:42.419 Awaish Kumar: But the thing is that, for the segment world.
128 00:17:42.590 ⇒ 00:17:46.260 Awaish Kumar: We restrict it to the limited number of
129 00:17:46.720 ⇒ 00:17:49.539 Awaish Kumar: sources that we can say, like, only the…
130 00:17:49.720 ⇒ 00:17:52.589 Awaish Kumar: These two or three sources should be there.
131 00:17:53.440 ⇒ 00:17:55.399 Awaish Kumar: In the, in the, in this school?
132 00:17:55.570 ⇒ 00:18:00.060 Awaish Kumar: Right? And, like, And end up…
133 00:18:01.170 ⇒ 00:18:04.360 Awaish Kumar: Yeah, we can define, for the mixed panel.
134 00:18:06.100 ⇒ 00:18:16.349 Awaish Kumar: So, yeah, my thing is that we want to limit it, the… the segments migration part, so it’s a lot of things here.
135 00:18:16.460 ⇒ 00:18:18.530 Awaish Kumar: And Yo.
136 00:18:18.530 ⇒ 00:18:26.419 Robert Tseng: I mean, Segment is propping up all these Basque webhooks, but aside from that, everything else is just marketing, just, like, ads, and .
137 00:18:27.320 ⇒ 00:18:27.960 Awaish Kumar: Yeah.
138 00:18:28.190 ⇒ 00:18:33.410 Robert Tseng: Yeah, so… point of what, this is… catalog here.
139 00:18:33.880 ⇒ 00:18:42.440 Robert Tseng: Yeah, so, I mean, yeah, I guess we’re, yeah, we’re not gonna migrate over this Basque stuff, fine, but, like, you know, everything else is pretty much just a marketing kind of thing anyway.
140 00:18:43.480 ⇒ 00:18:45.690 Awaish Kumar: Yeah… okay.
141 00:18:49.260 ⇒ 00:18:50.180 Awaish Kumar: Okay.
142 00:18:52.370 ⇒ 00:19:00.739 Awaish Kumar: But still, it helps us keep the scope a bit lower, so we can actually… achieve it.
143 00:19:01.580 ⇒ 00:19:07.879 Robert Tseng: Yeah, so I want… yeah, I think your, like, mindset should change, that it’s not like, okay, there’s gonna be a new inflow of, like.
144 00:19:08.010 ⇒ 00:19:14.840 Robert Tseng: you know, more engineering work requests for us to take on. I think it’s like, no, I think we can assume that they want to just rip this
145 00:19:15.430 ⇒ 00:19:17.459 Robert Tseng: Codex stuff, and just, like.
146 00:19:17.790 ⇒ 00:19:31.640 Robert Tseng: build as much as they can themselves. And they also have background engineers, we can just ask them to go and, like, contribute here and there, but we kind of have to just add… we have to just do the steering and, like, make sure that the scaffolding is there, like, you know, make sure that
147 00:19:31.670 ⇒ 00:19:44.059 Robert Tseng: Yeah, as they’re building into the data platform, that things don’t break, and that we can add the appropriate limits. I think that’s… that’s what our role ends up shifting to. So, like, ideally, you end up touching fewer, you know, like.
148 00:19:44.120 ⇒ 00:19:47.699 Robert Tseng: Tickets, because, like, you’re… you’re not really, like, having to
149 00:19:47.800 ⇒ 00:19:56.100 Robert Tseng: I mean, there’s gonna be some main… obviously, there’s maintenance for us, but, like, I don’t… I don’t want… like, have them do the migration. Like, even this stuff, like, you know.
150 00:19:56.870 ⇒ 00:20:04.470 Robert Tseng: whatever, moving some stuff, like, this from… from segment over into the infrastructure, like, have them do it. Like.
151 00:20:05.270 ⇒ 00:20:14.969 Awaish Kumar: I’m just concerned, because for… on Remo, we had a scope creep, like, we have always been asked to be responsible for the things that really we didn’t…
152 00:20:15.340 ⇒ 00:20:17.090 Awaish Kumar: We’re responsible for, so…
153 00:20:17.090 ⇒ 00:20:17.690 Robert Tseng: Yeah.
154 00:20:18.480 ⇒ 00:20:25.960 Awaish Kumar: I just wanted to define that so we… so that we have a fixed thing that we can actually deliver.
155 00:20:27.620 ⇒ 00:20:41.929 Robert Tseng: Yeah, I mean, like it or not, this client is just kinda gonna go off the rails. Like, they’re not gonna respect the scope creep. Like, that’s just not how… I mean, I’m doing my best to push back on it. The whole Remo thing, like, yes, I think there’s a lot there, but
156 00:20:42.020 ⇒ 00:20:55.360 Robert Tseng: and it’s… it’s inefficient, it causes the project to slow down, like, yes, all of those things are true, and Eden is not the ideal client. But I… yeah, so I… I think I’m just… whenever we’re solving a problem for one client, I’m always thinking about, like.
157 00:20:55.490 ⇒ 00:21:00.390 Robert Tseng: how does solving the problem here… is this problem worth solving? Because will we learn anything?
158 00:21:00.390 ⇒ 00:21:22.960 Robert Tseng: that we’ll be able to take to, like, other clients, right? And I think that’s always been my approach with Eden, that whatever we learn on Eden, I try to, like, apply it to another client. If it’s not something that I feel like would actually benefit our team, I just don’t even… I just say no to it. Like, I… I think so. So yeah, I think that’s… that’s kind of what I want to do. Like, yeah, I’m not trying to get you to just say yes to, like.
159 00:21:22.960 ⇒ 00:21:27.779 Robert Tseng: moving all these connectors or whatever. I think you gotta… yeah, I… yeah, so…
160 00:21:27.780 ⇒ 00:21:44.639 Robert Tseng: I’m… yeah, that’s not my intention here. So… I mean, if you think that I’m being too idealistic, and that’s, like, actually not how it works, then, like, you can let me know, but the way that I thought about it was that they could… that if they’re going to be messing around and pushing all these changes, just, like.
161 00:21:44.820 ⇒ 00:21:53.289 Robert Tseng: try to let them do that in a way that doesn’t involve us so much, and we’re just… just focusing on… on the governments and the DevOps part of it.
162 00:21:53.870 ⇒ 00:22:05.089 Awaish Kumar: So, so, okay, then I agree, with this approach that you are… you have written down. It’s mostly, like, helping them def… helping them define or define ourselves the…
163 00:22:05.350 ⇒ 00:22:15.160 Awaish Kumar: the best way to have defined configs, GitHub, PR processes, deployment to staging versus production for segment, and for.
164 00:22:15.160 ⇒ 00:22:30.009 Robert Tseng: Exactly. And whatever you build out there, I’ll be able to use that as a playbook for another client, right? So, like, yeah, that’s kind of why I view it as, like, still a good… a good, like, thing, because, you know, you’re gonna… if we walk into another situation.
165 00:22:30.130 ⇒ 00:22:36.270 Robert Tseng: Then we will have these art… like, we’ll have these resources already built out, we could just reuse it.
166 00:22:36.510 ⇒ 00:22:46.740 Awaish Kumar: Yeah, then I think we are on the same page, right? I was more thinking about actually doing a migration and from end to end, like, moving everything as an infrastructure as a code.
167 00:22:47.900 ⇒ 00:22:53.029 Robert Tseng: Yeah. I mean, we should tell them that that’s the direction that they’re… that seems like they’re headed in.
168 00:22:53.240 ⇒ 00:22:59.859 Robert Tseng: we’re not going to do the… we’re not going to do the migration for them. We’re just going to tell them, like, what they need to do to do it.
169 00:22:59.860 ⇒ 00:23:16.030 Robert Tseng: And if they want to try running it with their… with their agents as code, because they’re so confident that they can do it, go ahead and do it. We’ll… we’ll help you steer it. But, like, we’re not going to actually be on the hook for the migration. We’re just there for the governance. I think that’s how I want to… that’s how I want to position it.
170 00:23:16.580 ⇒ 00:23:18.740 Awaish Kumar: Okay, yeah, then I agree with you.
171 00:23:19.500 ⇒ 00:23:20.140 Robert Tseng: Okay.
172 00:23:21.080 ⇒ 00:23:32.060 Robert Tseng: Cool. Yeah, and then if it comes back, and they’re like, oh, like, why don’t you help us, like, get over the line? It’s like, well, you know what? We already have something that works, and we maintain it. Like, we’re not trying to migrate it. It doesn’t make any difference, like, for us.
173 00:23:32.060 ⇒ 00:23:49.700 Robert Tseng: you just want to have the perception that, like, you’re contributing to this, you’re simplifying a stack, and realizing it’s not that easy. So, like, you know, we… our… we are not going to go and do this migration, like, it just seems like it would… I mean, I would charge a bunch more for it, and
174 00:23:49.700 ⇒ 00:24:03.169 Robert Tseng: yeah, if you want us to just maintain what we have and, like, do the governance to, help you, like, go and push code, then we will do that. That’s basically how I want to protect the team.
175 00:24:03.650 ⇒ 00:24:04.520 Awaish Kumar: Okay, okay.
176 00:24:04.840 ⇒ 00:24:08.590 Robert Tseng: Yeah. Okay, cool, so I think we’re on the same page, yeah, so if you don’t mind…
177 00:24:08.590 ⇒ 00:24:26.369 Robert Tseng: just kind of, yeah, two things, just going through this doc, making sure… I mean, it sounds like I missed the mark on a couple things here, so if you could go and make the edits, even if you don’t have the wording, at least just kind of strike out what you think is not necessary. I’m thinking even, like, these two weeks need to be broken out, so I have some, like, thoughts there.
178 00:24:26.780 ⇒ 00:24:27.180 Awaish Kumar: Okay.
179 00:24:27.180 ⇒ 00:24:33.799 Robert Tseng: I basically missed the whole infrastructure kind of design part, or like, kind of deciding what those are, so having your
180 00:24:34.160 ⇒ 00:24:40.459 Robert Tseng: input there would be helpful. And then, yeah, and this is just, like, one slide that I tagged you in.
181 00:24:40.460 ⇒ 00:24:46.040 Awaish Kumar: So I have created something for this, I don’t know how granular you want, so can I show you something?
182 00:24:46.460 ⇒ 00:24:48.920 Robert Tseng: Yeah, yeah, okay, let’s just… yeah, sure.
183 00:24:50.830 ⇒ 00:24:55.760 Awaish Kumar: So, actually, I… Where’s that?
184 00:24:57.190 ⇒ 00:24:58.030 Awaish Kumar: Okay.
185 00:25:03.640 ⇒ 00:25:09.839 Awaish Kumar: So, basically, I’m using Mermaid to, like, quickly come up with something, otherwise.
186 00:25:09.840 ⇒ 00:25:10.180 Robert Tseng: Great.
187 00:25:10.360 ⇒ 00:25:12.039 Awaish Kumar: It’ll take a lot of time.
188 00:25:13.220 ⇒ 00:25:14.610 Robert Tseng: Yeah, Figma sucks.
189 00:25:16.310 ⇒ 00:25:21.149 Awaish Kumar: So… This is a… this is one of the images.
190 00:25:21.790 ⇒ 00:25:27.550 Awaish Kumar: Although it doesn’t look nice, but it captures everything, like, the lineage.
191 00:25:28.540 ⇒ 00:25:30.770 Awaish Kumar: From… these are Mark’s table.
192 00:25:31.480 ⇒ 00:25:34.020 Robert Tseng: Yeah, just send that to me, let me try to make it nicer.
193 00:25:34.290 ⇒ 00:25:39.329 Awaish Kumar: So this is one thing, and another one, I have a version of it, which is with
194 00:25:39.830 ⇒ 00:25:45.930 Awaish Kumar: Nicer, like, but it encapsulates the… the raw…
195 00:25:46.300 ⇒ 00:25:53.150 Robert Tseng: Have you tried using the diagram builder that, like, B and Utan built out? I haven’t used it, because I’m not confident that it’s good, but…
196 00:25:53.620 ⇒ 00:26:00.219 Awaish Kumar: Yeah, I haven’t tried it, but I will… I was looking for that while working on it. So this is the…
197 00:26:00.410 ⇒ 00:26:03.309 Awaish Kumar: This is, like, the final March table.
198 00:26:03.690 ⇒ 00:26:07.020 Robert Tseng: Great, that’s… that looks clean. I think that… I would just… I’d rather do that, so…
199 00:26:07.270 ⇒ 00:26:16.089 Awaish Kumar: The only thing is that they don’t know how many raw tables we are bringing the data from to actually make this one table, for example.
200 00:26:16.370 ⇒ 00:26:25.560 Robert Tseng: I think what would be easy to do is, once I get that, you just tell me, like, what number, I’ll just put the number on top of it, you know? Rather than having, like, you know, 20 bars, I will just put, like.
201 00:26:25.850 ⇒ 00:26:28.889 Robert Tseng: didn’t know us about tables. There’s, like, 20 tables here, or whatever, so…
202 00:26:29.120 ⇒ 00:26:31.840 Awaish Kumar: In the raw, there are 60 tables, but for…
203 00:26:31.840 ⇒ 00:26:33.959 Robert Tseng: Yeah, so I just… you subbed me on that.
204 00:26:34.630 ⇒ 00:26:43.339 Awaish Kumar: Okay, I will send you… I will add this to slide, and maybe keep the other one in the slide, so if you… if they want to more… go more granular, you can use that.
205 00:26:43.730 ⇒ 00:26:45.370 Robert Tseng: Okay, appreciate it.
206 00:26:45.750 ⇒ 00:26:46.720 Awaish Kumar: Okay, thanks.
207 00:26:47.750 ⇒ 00:26:54.529 Robert Tseng: Cool, yeah, I think that should be good. So hopefully that was helpful, you know, I’m trying to,
208 00:26:55.440 ⇒ 00:27:07.369 Robert Tseng: Yeah, I’m… I’m just… I’m just trying to… trying to help on where I feel like is important on the eating side. So, yeah, let me know, let me know where else, but I would like to send them that doc later today. Okay.
209 00:27:08.070 ⇒ 00:27:12.920 Robert Tseng: And that’ll probably kick off a couple… couple things from there. Yeah.
210 00:27:13.680 ⇒ 00:27:14.840 Awaish Kumar: Okay, sure, thank you.
211 00:27:14.840 ⇒ 00:27:16.320 Robert Tseng: Okay, talk to you later.
212 00:27:16.570 ⇒ 00:27:18.020 Robert Tseng: Thanks. Bye.