Meeting Title: Brainforge-Engineering-Roadmap Date: 2024-09-27 Meeting participants: Nicolas Sucari, Roy Christian Piñon, Uttam Kumaran, Patrick Trainer
WEBVTT
1 00:00:08.590 ⇒ 00:00:11.284 Uttam Kumaran: Hey, guys, just gonna go just gonna get a coffee. But
2 00:00:11.880 ⇒ 00:00:14.559 Uttam Kumaran: let’s go off. Video for a sec.
3 00:00:18.270 ⇒ 00:00:19.559 Uttam Kumaran: Cool. That was great.
4 00:00:23.310 ⇒ 00:00:24.210 Uttam Kumaran: Yeah.
5 00:00:24.610 ⇒ 00:00:25.370 Patrick Trainer: That one will.
6 00:00:28.150 ⇒ 00:00:29.461 Uttam Kumaran: What do you guys think of
7 00:00:30.360 ⇒ 00:00:32.650 Uttam Kumaran: of all the stuff like on the sales side?
8 00:00:34.610 ⇒ 00:00:35.209 Patrick Trainer: Ready to.
9 00:00:35.210 ⇒ 00:00:36.210 Nicolas Sucari: Really interesting.
10 00:00:36.210 ⇒ 00:00:37.290 Patrick Trainer: Start, selling.
11 00:00:37.530 ⇒ 00:00:37.875 Uttam Kumaran: Yeah.
12 00:00:39.510 ⇒ 00:00:40.363 Uttam Kumaran: me, too.
13 00:00:41.780 ⇒ 00:00:44.870 Patrick Trainer: Yeah, it all sounds good. So
14 00:00:45.460 ⇒ 00:00:47.010 Patrick Trainer: let’s make some money.
15 00:00:54.770 ⇒ 00:00:58.139 Uttam Kumaran: Yeah, you can go ahead. I have my air pods in, but I’m listening.
16 00:00:58.150 ⇒ 00:00:59.969 Uttam Kumaran: Do you want to tee up whatever.
17 00:01:03.080 ⇒ 00:01:03.420 Patrick Trainer: Yes.
18 00:01:03.420 ⇒ 00:01:11.209 Uttam Kumaran: Yeah, I just wanna maximize this time. I’m sure there’ll be a bunch of follow ups. But yeah, let’s let’s get into it.
19 00:01:11.541 ⇒ 00:01:13.858 Patrick Trainer: I guess with it I’ve got like
20 00:01:14.560 ⇒ 00:01:18.719 Patrick Trainer: like some visual assets, too. So I’m not sure if you wanna
21 00:01:19.700 ⇒ 00:01:21.210 Patrick Trainer: wait there. But
22 00:01:21.220 ⇒ 00:01:29.650 Patrick Trainer: basically. So the way that I’ve subjugated this out is like, we’ve got strategy implementation
23 00:01:30.146 ⇒ 00:01:44.820 Patrick Trainer: and then kind of like a client roadmap of like how we understand data problems of people or of of clients. And then like back that into like, what does our infrastructure need to support?
24 00:01:44.990 ⇒ 00:01:48.104 Patrick Trainer: And then I’ve got some high level
25 00:01:48.600 ⇒ 00:01:54.819 Patrick Trainer: and a little bit of a lower level diagram of each component that I see
26 00:01:54.880 ⇒ 00:01:59.139 Patrick Trainer: in terms of like an all encompassing data platform
27 00:01:59.917 ⇒ 00:02:05.609 Patrick Trainer: and so with that like, I mean, don’t think of this as like
28 00:02:05.630 ⇒ 00:02:12.409 Patrick Trainer: written in stone. For much, I’d say, this is like the probably like the second or 3rd pass
29 00:02:12.866 ⇒ 00:02:19.109 Patrick Trainer: that we’ve made now on what we want our infrastructure to really look like.
30 00:02:19.350 ⇒ 00:02:24.150 Patrick Trainer: And so with that, let me.
31 00:02:25.340 ⇒ 00:02:26.249 Patrick Trainer: where am I going?
32 00:02:28.860 ⇒ 00:02:30.940 Patrick Trainer: Let me, where did Zoom go?
33 00:02:33.492 ⇒ 00:02:35.480 Patrick Trainer: Share the screen?
34 00:02:39.260 ⇒ 00:02:40.210 Patrick Trainer: Okay.
35 00:02:41.210 ⇒ 00:02:43.910 Patrick Trainer: does it switch to? Oh, no, it doesn’t.
36 00:02:45.320 ⇒ 00:02:47.890 Patrick Trainer: Oh, yeah, okay. You can see
37 00:02:48.850 ⇒ 00:02:50.500 Patrick Trainer: Vs code right?
38 00:02:50.830 ⇒ 00:02:53.210 Patrick Trainer: Or are you looking at my slack.
39 00:02:55.040 ⇒ 00:02:56.220 Nicolas Sucari: Obvious code. That’s okay.
40 00:02:56.220 ⇒ 00:02:57.320 Patrick Trainer: Okay, cool.
41 00:02:57.380 ⇒ 00:02:58.619 Patrick Trainer: cool, cool, cool.
42 00:02:58.930 ⇒ 00:03:00.110 Patrick Trainer: So
43 00:03:00.290 ⇒ 00:03:08.873 Patrick Trainer: let me actually 1st go into sheets here. So basically, this is what I was going through
44 00:03:10.330 ⇒ 00:03:11.580 Patrick Trainer: just before
45 00:03:12.074 ⇒ 00:03:20.350 Patrick Trainer: in kind of like our problem matrix of like what we need to think about. And so we’ve got multiple projects.
46 00:03:20.400 ⇒ 00:03:21.949 Patrick Trainer: we need to
47 00:03:22.550 ⇒ 00:03:24.810 Patrick Trainer: handle fragmented tools.
48 00:03:25.050 ⇒ 00:03:26.210 Patrick Trainer: growth
49 00:03:26.310 ⇒ 00:03:28.190 Patrick Trainer: assets, all of that.
50 00:03:28.370 ⇒ 00:03:43.490 Patrick Trainer: And basically, we have all like these benefits of what we’re looking for. So it’s like, improve our resource utilization resources just not compute, but also in terms of like engineering hours, and just kind of like headspace.
51 00:03:43.570 ⇒ 00:03:52.229 Patrick Trainer: And then having that consistent policy so like, think like templates and whatnot of being able to do this repeatedly.
52 00:03:52.602 ⇒ 00:03:56.970 Patrick Trainer: And then we also need to be able to see into it, manage it.
53 00:03:57.475 ⇒ 00:04:07.480 Patrick Trainer: have good understanding of like, what’s being governed and then have just like our our technical alignment with with business.
54 00:04:07.910 ⇒ 00:04:09.280 Patrick Trainer: and then
55 00:04:09.400 ⇒ 00:04:15.326 Patrick Trainer: ensuring that we have like quality standards across all of our clients, I think.
56 00:04:16.839 ⇒ 00:04:19.800 Patrick Trainer: I mean, it’s basically like templatizing.
57 00:04:19.930 ⇒ 00:04:25.529 Patrick Trainer: like, if we have one ideal project being able to repeatedly.
58 00:04:26.046 ⇒ 00:04:30.930 Patrick Trainer: Implement that across all of the clients is going to be like
59 00:04:31.470 ⇒ 00:04:47.449 Patrick Trainer: in my eyes. What makes or breaks the company? If we’re if we have one good client with a project or one good project for a client like that’s great. But if we can’t replicate that across the stack or across the the client base. Then
60 00:04:47.720 ⇒ 00:04:52.559 Patrick Trainer: people are going to be pissed off, and everything. Nothing will be predictable
61 00:04:53.104 ⇒ 00:04:59.460 Patrick Trainer: and so like. That’s super key in being able to like, replicate that experience
62 00:04:59.480 ⇒ 00:05:04.399 Patrick Trainer: every single time. And so in doing that, we need to
63 00:05:04.430 ⇒ 00:05:09.189 Patrick Trainer: like engineer this platform to be able to to handle those sorts of things.
64 00:05:09.708 ⇒ 00:05:16.541 Patrick Trainer: And then I’ve got a bit of like risk assessment matrix there of
65 00:05:17.210 ⇒ 00:05:22.820 Patrick Trainer: common issues that we need to think about when building this this platform
66 00:05:23.568 ⇒ 00:05:29.631 Patrick Trainer: and so in those we’re thinking about like security breaches, scalability like, how well does it?
67 00:05:30.130 ⇒ 00:05:32.929 Patrick Trainer: like, how well can we use it?
68 00:05:34.000 ⇒ 00:05:42.779 Patrick Trainer: user adoption. That’s more of like us in our engineers like, how easy are these tools to to use?
69 00:05:43.177 ⇒ 00:05:53.539 Patrick Trainer: Can we integrate other tools into it? That’s super important, especially as we’re wanting to stay on kind of like the the cutting edge of things here.
70 00:05:53.750 ⇒ 00:05:56.900 Patrick Trainer: as well as being
71 00:05:56.930 ⇒ 00:06:07.330 Patrick Trainer: compliant with all of our regulations. That’s I mean, incredibly important, just as especially as we grow like companies always love to ask for
72 00:06:07.400 ⇒ 00:06:17.861 Patrick Trainer: like sock 2, and whether you think it’s like it’s bullshit or not like it kind of is, but it’s like that stamp of approval that shows some like
73 00:06:18.860 ⇒ 00:06:22.519 Patrick Trainer: like we we’ve at least looked into this a little bit.
74 00:06:23.206 ⇒ 00:06:29.000 Patrick Trainer: And so going forward with all of this in mind. What we’re wanting to
75 00:06:29.090 ⇒ 00:06:32.189 Patrick Trainer: show here is, and this gets like
76 00:06:32.310 ⇒ 00:06:35.100 Patrick Trainer: pretty deep. So bear with me.
77 00:06:35.280 ⇒ 00:06:46.270 Patrick Trainer: we’ve got running through our platform architecture. This. This relates to that 1st box over here in our problem matrix
78 00:06:46.410 ⇒ 00:06:49.989 Patrick Trainer: of like managing multiple data projects, right?
79 00:06:50.150 ⇒ 00:06:51.220 Patrick Trainer: And so
80 00:06:51.870 ⇒ 00:06:53.100 Patrick Trainer: as they grow
81 00:06:54.140 ⇒ 00:06:55.030 Patrick Trainer: we’re
82 00:06:56.130 ⇒ 00:07:16.760 Patrick Trainer: basically already describe that. But we have what we’re wanting to do. So we have these different tenants. Those need to be isolated from each other. And we need to be able to share services in between them. So it’s kind of like our central repository theory with branched out of there
83 00:07:17.239 ⇒ 00:07:37.160 Patrick Trainer: and making sure that like that’s simple to manage and that goes into like the templates in in doing all of that, but then also like being able to manage all of our like logs. Authentication to all of these different services in a way that’s not gonna like, break our brains.
84 00:07:37.650 ⇒ 00:07:46.409 Patrick Trainer: And from there, where this comes like, I, what I see is like the the middle piece is this like centralized control plane?
85 00:07:46.430 ⇒ 00:07:48.609 Patrick Trainer: So if you think like.
86 00:07:48.900 ⇒ 00:07:51.699 Patrick Trainer: like in in Star Trek, when they’re
87 00:07:51.950 ⇒ 00:08:00.140 Patrick Trainer: managing all of their different like fleets. They’ve got one screen where they’re kind of like moving everything together and like wargaming in that sense.
88 00:08:00.160 ⇒ 00:08:03.220 Patrick Trainer: That’s how I envision the data platform
89 00:08:03.310 ⇒ 00:08:18.120 Patrick Trainer: in that, like you can see the overall health of all projects at the same time. So that’s that like, that’s like this here, that like unified, unified dashboard, essentially.
90 00:08:18.130 ⇒ 00:08:19.360 Patrick Trainer: And that’s
91 00:08:19.890 ⇒ 00:08:22.549 Patrick Trainer: brain forge specific and not
92 00:08:22.590 ⇒ 00:08:40.567 Patrick Trainer: not so in. In all of this, this is what I’m talking about. And it’s brand for specific, and that all of the projects are like clients, right? And so we should be able to have, like a high level overview of all of our clients, all of the resources that are using and then be able to zoom in and see those metrics.
93 00:08:40.840 ⇒ 00:08:50.420 Nicolas Sucari: And and sorry question. There! How do you imagine like that dashboard, or on which tool do you imagine having that dashboard working it should be.
94 00:08:50.800 ⇒ 00:08:51.730 Nicolas Sucari: I don’t know.
95 00:08:52.890 ⇒ 00:08:54.849 Patrick Trainer: So so I
96 00:08:55.500 ⇒ 00:08:57.420 Patrick Trainer: I would I
97 00:08:57.950 ⇒ 00:09:05.540 Patrick Trainer: I’d phrase that in a way of like, let’s not think about like, what specific tool do we use to create a dashboard.
98 00:09:05.580 ⇒ 00:09:07.279 Patrick Trainer: but more so like
99 00:09:07.450 ⇒ 00:09:08.450 Patrick Trainer: like, because a lot of.
100 00:09:08.450 ⇒ 00:09:08.830 Nicolas Sucari: Or.
101 00:09:08.830 ⇒ 00:09:10.949 Patrick Trainer: To be like custom, code.
102 00:09:10.950 ⇒ 00:09:11.590 Uttam Kumaran: Yeah.
103 00:09:11.590 ⇒ 00:09:12.100 Patrick Trainer: And.
104 00:09:12.100 ⇒ 00:09:12.710 Nicolas Sucari: Okay.
105 00:09:12.710 ⇒ 00:09:13.800 Patrick Trainer: Doing it that way.
106 00:09:13.800 ⇒ 00:09:18.799 Uttam Kumaran: You’re thinking something like some something like retool, or like an internal basically like control plane. Okay.
107 00:09:19.030 ⇒ 00:09:19.340 Patrick Trainer: Exactly.
108 00:09:19.340 ⇒ 00:09:19.630 Nicolas Sucari: Okay.
109 00:09:19.630 ⇒ 00:09:22.839 Patrick Trainer: Yeah, yeah, less, less. So like.
110 00:09:23.530 ⇒ 00:09:25.859 Patrick Trainer: let’s use looker to.
111 00:09:25.860 ⇒ 00:09:31.160 Nicolas Sucari: No, no, that’s fine. That’s fine. I’m trying to. Yeah. I’m trying to picture how we are. Gonna
112 00:09:31.390 ⇒ 00:09:43.449 Nicolas Sucari: because if we are like bringing different sources in, and a lot of different integrations. The data like having a place where we can check, like, yeah, health of everything.
113 00:09:43.450 ⇒ 00:09:46.679 Patrick Trainer: I kind of. I kind of get to it down
114 00:09:47.130 ⇒ 00:09:49.750 Patrick Trainer: like I’ve I’ve written a lot
115 00:09:50.220 ⇒ 00:09:52.903 Patrick Trainer: here as we go.
116 00:09:53.970 ⇒ 00:09:59.310 Patrick Trainer: but as it comes. It’s like we basically
117 00:09:59.450 ⇒ 00:10:04.586 Patrick Trainer: are designing a service, but like a service for ourselves.
118 00:10:05.100 ⇒ 00:10:05.400 Nicolas Sucari: Okay.
119 00:10:05.400 ⇒ 00:10:18.890 Patrick Trainer: And like a product for ourselves. And so for that, to be able to just like work and get off the ground, it’s going to need to be Api driven. And like, we need to have really like an Api 1st mentality
120 00:10:18.930 ⇒ 00:10:22.430 Patrick Trainer: of like, how are we
121 00:10:22.910 ⇒ 00:10:27.320 Patrick Trainer: like building this stuff so like right now, like we’re kind of
122 00:10:27.480 ⇒ 00:10:29.760 Patrick Trainer: shooting from the hip in every.
123 00:10:29.760 ⇒ 00:10:30.569 Nicolas Sucari: Okay, cool. Like.
124 00:10:30.570 ⇒ 00:10:34.130 Patrick Trainer: Client. And we’re just like building to those specific needs.
125 00:10:34.570 ⇒ 00:10:40.119 Patrick Trainer: Here is like, we’re shifting that paradigm. And like we’re.
126 00:10:40.500 ⇒ 00:10:48.779 Patrick Trainer: I mean, honestly, not really thinking about a client or any specific client at all. Here we’re thinking of like
127 00:10:49.080 ⇒ 00:10:55.880 Patrick Trainer: brain forge is is the client. And these are the problems that brain forge has. How do we solve those?
128 00:10:55.900 ⇒ 00:10:59.570 Patrick Trainer: And so that’s kind of like where this is coming from?
129 00:10:59.910 ⇒ 00:11:06.620 Patrick Trainer: And so let’s see, can I? I can’t collapse these. So this will have.
130 00:11:06.760 ⇒ 00:11:08.599 Patrick Trainer: You can see all of this.
131 00:11:08.600 ⇒ 00:11:09.210 Uttam Kumaran: Yeah.
132 00:11:09.370 ⇒ 00:11:12.080 Patrick Trainer: Actually right? Okay, so
133 00:11:12.220 ⇒ 00:11:17.670 Patrick Trainer: this is just like a higher level overview of everything that’s in here.
134 00:11:17.840 ⇒ 00:11:22.630 Patrick Trainer: And so we’ve got our architecture, which we’re kind of going through here.
135 00:11:22.630 ⇒ 00:11:28.210 Uttam Kumaran: You may not need the the actual raw code on the left, like we could probably just look at the
136 00:11:28.810 ⇒ 00:11:31.230 Patrick Trainer: I mean, yeah, that’s that’s a good point.
137 00:11:33.220 ⇒ 00:11:34.390 Uttam Kumaran: Nice. Okay. Cool.
138 00:11:36.245 ⇒ 00:11:37.140 Patrick Trainer: Shit.
139 00:11:52.450 ⇒ 00:11:53.710 Nicolas Sucari: Maybe we need.
140 00:11:54.030 ⇒ 00:11:55.389 Uttam Kumaran: Maybe we do need both. I don’t know.
141 00:11:55.390 ⇒ 00:11:56.090 Patrick Trainer: Maybe we did.
142 00:11:56.090 ⇒ 00:11:57.099 Nicolas Sucari: Here, I’ll just make that one.
143 00:11:57.100 ⇒ 00:11:58.160 Uttam Kumaran: It’s something small.
144 00:11:58.710 ⇒ 00:12:03.499 Patrick Trainer: To keep the outline up. Yeah, you need the code. It’s I don’t know. It’s a pain.
145 00:12:06.060 ⇒ 00:12:08.170 Patrick Trainer: But I think, yeah, this this works.
146 00:12:08.210 ⇒ 00:12:09.539 Patrick Trainer: So we’ve got
147 00:12:09.690 ⇒ 00:12:11.580 Patrick Trainer: going through. We’ve got like.
148 00:12:13.810 ⇒ 00:12:15.332 Patrick Trainer: God damn it.
149 00:12:17.010 ⇒ 00:12:34.610 Patrick Trainer: we’ve got this like centralized control plan. So this goes into our like Api design 1st Api 1st design. So like, we want to be able to essentially like programmatically control all of our projects, or at least be able to control like into those projects. So
150 00:12:35.130 ⇒ 00:12:38.130 Patrick Trainer: instead, like, think of like, if we’re running like
151 00:12:38.160 ⇒ 00:12:47.330 Patrick Trainer: Github, actions, like all of those, are based off of like web hooks events, or like triggering events into those to kick things off
152 00:12:47.930 ⇒ 00:12:55.750 Patrick Trainer: all of those like if you think of like, if you have a a page here and we have a button that is like.
153 00:12:55.860 ⇒ 00:13:02.210 Patrick Trainer: run this workflow, you click that button. It runs the workflow. That’s the kind of like stuff that I’m thinking of.
154 00:13:02.788 ⇒ 00:13:09.641 Patrick Trainer: And then that infrastructure needs to be like scalable especially like as as volumes grow.
155 00:13:10.350 ⇒ 00:13:13.400 Patrick Trainer: all all of that. And so
156 00:13:13.680 ⇒ 00:13:23.629 Patrick Trainer: gonna already have this. But this is going to go hand in hand with like client stuff, but also be in in our purview as well.
157 00:13:24.189 ⇒ 00:13:33.380 Patrick Trainer: And so we come into data governance. What does that mean? We need to understand like what’s out there
158 00:13:33.420 ⇒ 00:13:39.030 Patrick Trainer: and be able, like, I’d like to be able to see this across, like all projects
159 00:13:39.160 ⇒ 00:13:41.859 Patrick Trainer: in one sort of view.
160 00:13:42.524 ⇒ 00:13:44.486 Patrick Trainer: As opposed to
161 00:13:45.410 ⇒ 00:13:54.902 Patrick Trainer: like having to go into each and individual project and be like, okay, this is here. This is here and not be able to like cross. Compare those.
162 00:13:55.370 ⇒ 00:14:04.249 Patrick Trainer: that’s that’s kind of how I’m thinking about it, like some of the questions that, like like Brian was asking yesterday of like having
163 00:14:04.760 ⇒ 00:14:08.880 Patrick Trainer: Gpt have the context of like, oh, we’ve used
164 00:14:09.350 ⇒ 00:14:10.440 Patrick Trainer: this
165 00:14:10.930 ⇒ 00:14:17.339 Patrick Trainer: SQL function over here this. It’s I see this in a similar lens of like
166 00:14:17.730 ⇒ 00:14:25.549 Patrick Trainer: this is what the structure of like this catalog looks like. And this catalog looks like and like providing that context back and forth.
167 00:14:26.240 ⇒ 00:14:30.979 Nicolas Sucari: But we can do that with like different clients, like having the same.
168 00:14:31.190 ⇒ 00:14:34.619 Nicolas Sucari: It’s not sharing the data between them, right? But it’s.
169 00:14:34.620 ⇒ 00:14:35.340 Patrick Trainer: Right, right, yeah.
170 00:14:35.340 ⇒ 00:14:40.589 Nicolas Sucari: Using one client’s data to start building the structure or
171 00:14:40.600 ⇒ 00:14:43.919 Nicolas Sucari: for another one like, is that possible? Or it’s kind of like.
172 00:14:43.920 ⇒ 00:14:45.019 Patrick Trainer: Right? Right? It’s so.
173 00:14:45.020 ⇒ 00:14:45.760 Nicolas Sucari: Don’t mess up.
174 00:14:45.760 ⇒ 00:14:47.700 Patrick Trainer: Again. If you think like
175 00:14:47.830 ⇒ 00:14:53.050 Patrick Trainer: that, this is for us like we’re we like. There’s
176 00:14:53.200 ⇒ 00:14:57.930 Patrick Trainer: Snowflake Apis that, like we would go and see that information, schema.
177 00:14:58.040 ⇒ 00:15:04.339 Patrick Trainer: that information schema is the same across each client or.
178 00:15:04.340 ⇒ 00:15:13.600 Nicolas Sucari: What we are. We are not sharing the data. We’re not sharing the data or the context. We’re we’re just sharing how we are structuring different things. And
179 00:15:14.036 ⇒ 00:15:16.460 Nicolas Sucari: for each of the clients. Right? Correct.
180 00:15:17.186 ⇒ 00:15:21.589 Patrick Trainer: And so like with that like, what’s a concrete example of like.
181 00:15:21.690 ⇒ 00:15:23.420 Patrick Trainer: how we want to
182 00:15:23.790 ⇒ 00:15:30.130 Patrick Trainer: understand what’s happening in each client’s deployment. And that’s going to come with like access control
183 00:15:30.633 ⇒ 00:15:39.296 Patrick Trainer: especially like in terms of us. So like, if we have specific engineers working on one project versus another project.
184 00:15:40.340 ⇒ 00:15:55.719 Patrick Trainer: we want to be able to like ensure in a readily like quick fashion, that, like things, are the way that they are, but then also be able to like provision those those grants or revoking those grants
185 00:15:55.800 ⇒ 00:15:57.530 Patrick Trainer: in a like
186 00:15:58.910 ⇒ 00:16:05.069 Patrick Trainer: in a simple way, so that we’re not having to like go in and like, basically just like.
187 00:16:05.190 ⇒ 00:16:11.499 Patrick Trainer: have that gut feeling that it’s okay, like, we know definitively that things are good.
188 00:16:13.700 ⇒ 00:16:17.260 Patrick Trainer: we also go into like defining
189 00:16:17.280 ⇒ 00:16:19.329 Patrick Trainer: like, what is data quality.
190 00:16:19.340 ⇒ 00:16:24.800 Patrick Trainer: We’ve talked a lot about this before. But then there’s also
191 00:16:26.490 ⇒ 00:16:28.410 Patrick Trainer: like other ways that
192 00:16:28.870 ⇒ 00:16:32.740 Patrick Trainer: we haven’t really discussed. I
193 00:16:33.310 ⇒ 00:16:38.520 Patrick Trainer: Udom on Linkedin the other day. I saw it like with with Toby
194 00:16:39.092 ⇒ 00:16:45.599 Patrick Trainer: the SQL. Mesh, Guy, you’re talking about unit testing. And so like
195 00:16:45.970 ⇒ 00:16:48.720 Patrick Trainer: that is one thing that we’re
196 00:16:49.470 ⇒ 00:17:07.490 Patrick Trainer: not really doing now, like we have some tests. But like most of those tests, correspond to like actual client data. And how like that client data is is controlled. But like, how are we testing the quality of like our logic in implementing that for them?
197 00:17:07.896 ⇒ 00:17:15.949 Patrick Trainer: That’s going into like Ryan’s Chat Gpt integration with that. But then also thinking of that in like a broader
198 00:17:15.960 ⇒ 00:17:22.860 Patrick Trainer: like, are we doing the correct thing and so coming up with ways to
199 00:17:23.500 ⇒ 00:17:27.670 Patrick Trainer: check and ensure that quality is is
200 00:17:28.359 ⇒ 00:17:30.839 Patrick Trainer: pertinent and top of mind. And so
201 00:17:30.880 ⇒ 00:17:36.329 Patrick Trainer: we have basically like our rules, which are like the key, like components of that.
202 00:17:36.790 ⇒ 00:17:39.380 Patrick Trainer: And then the actual monitoring
203 00:17:39.470 ⇒ 00:17:45.779 Patrick Trainer: of that of like what’s actually happening. What’s going on like is this making sense?
204 00:17:46.420 ⇒ 00:17:50.169 Patrick Trainer: And then we have the the remediation piece of it, too.
205 00:17:50.230 ⇒ 00:17:51.240 Patrick Trainer: of like.
206 00:17:51.460 ⇒ 00:17:56.079 Patrick Trainer: what do we do when something goes wrong or like, how how do we fix this.
207 00:17:56.230 ⇒ 00:17:56.690 Uttam Kumaran: Yeah.
208 00:17:57.860 ⇒ 00:17:59.790 Patrick Trainer: and that this
209 00:17:59.930 ⇒ 00:18:04.430 Patrick Trainer: dovetails right into monitoring and observability.
210 00:18:04.580 ⇒ 00:18:08.990 Patrick Trainer: And so this is going to be like our pipeline, so like which
211 00:18:09.010 ⇒ 00:18:13.042 Patrick Trainer: github pipelines are running, which ones are are failing
212 00:18:14.150 ⇒ 00:18:22.160 Patrick Trainer: even going into like the stuff that, like Miguel’s doing, even into our sales. Pipelines like we need to
213 00:18:23.270 ⇒ 00:18:29.520 Patrick Trainer: be able to like, understand, like one what pipes are there, and 2 if they’re flowing.
214 00:18:29.550 ⇒ 00:18:41.910 Patrick Trainer: and 3 if they’re flowing correctly. But then we also like need to go further into like, not just like internal pipelines for our internal operations, but like a step
215 00:18:42.300 ⇒ 00:19:01.830 Patrick Trainer: deeper into our clients, pipelines and client operations there. And that’s when we come with like Sla’s and slos, and so defining those is going to be important as well but then also for us and for our clients, too, is
216 00:19:01.900 ⇒ 00:19:09.590 Patrick Trainer: being able to track our resources. So there, there’s, I mean, a huge cost component to it, especially like on the client side.
217 00:19:09.650 ⇒ 00:19:12.410 Patrick Trainer: but being able to
218 00:19:12.500 ⇒ 00:19:18.060 Patrick Trainer: understand, that is going to tie into our slas and slos.
219 00:19:18.180 ⇒ 00:19:18.800 Patrick Trainer: Okay,
220 00:19:20.190 ⇒ 00:19:24.509 Patrick Trainer: coming into, I mean a little further into
221 00:19:24.680 ⇒ 00:19:27.267 Patrick Trainer: incidents and alerting.
222 00:19:28.150 ⇒ 00:19:30.970 Patrick Trainer: we kind of tried this with
223 00:19:31.830 ⇒ 00:19:33.571 Patrick Trainer: what you’re gonna call it,
224 00:19:34.110 ⇒ 00:19:44.929 Patrick Trainer: like element or elementary work to a degree. But I think there’s a lot more room for improvement there. It. It was.
225 00:19:44.930 ⇒ 00:19:47.549 Uttam Kumaran: I think we I mean look, I wanted to just
226 00:19:47.780 ⇒ 00:19:50.820 Uttam Kumaran: go a mile deep and then see what happened.
227 00:19:50.820 ⇒ 00:19:51.610 Patrick Trainer: Right.
228 00:19:51.610 ⇒ 00:19:56.100 Uttam Kumaran: Fail like as fast as possible, which I unfortunately it was not that fast.
229 00:19:56.150 ⇒ 00:20:10.239 Uttam Kumaran: but, like kind of get a sense for the appetite right and across like me, you and Ryan get a sense for like, where did this like, how does this gonna flush out? So I thought, I think it’s I don’t worry about it, not working out perfectly like I’m glad.
230 00:20:10.240 ⇒ 00:20:20.009 Patrick Trainer: Yeah. Oh, yeah, it’s I’m I’m not either. And I think it gave us like a lot of insight into like one for our clients like what matters and like what?
231 00:20:20.250 ⇒ 00:20:24.159 Patrick Trainer: And almost in a sense of like, where’s the 1st step?
232 00:20:24.572 ⇒ 00:20:33.359 Patrick Trainer: And like like what? Step? One step, 2, step 3, I think elementary was like, Step 4, 5. And really, when we just need like step one and 2
233 00:20:33.842 ⇒ 00:20:38.300 Patrick Trainer: but then, like we’ve got alerting, I think a lot of this is.
234 00:20:38.320 ⇒ 00:20:43.709 Patrick Trainer: and I mean a lot of this is too further down the road. But for the
235 00:20:43.730 ⇒ 00:20:49.982 Patrick Trainer: sake of this talk I’m just like walking through kind of like all of the different rungs of
236 00:20:50.380 ⇒ 00:20:51.640 Patrick Trainer: where we go.
237 00:20:52.119 ⇒ 00:21:01.129 Patrick Trainer: And so this includes like getting wrong books, so like when errors happen when we have alerts, how do we remediate them?
238 00:21:01.715 ⇒ 00:21:05.169 Patrick Trainer: Are they easy to use, and
239 00:21:05.420 ⇒ 00:21:11.279 Patrick Trainer: are they auditable in the sense that like, how do we communicate back to our clients like.
240 00:21:11.490 ⇒ 00:21:26.449 Patrick Trainer: oh, we found this issue, or this happened, and we missed like these insights for you like, how do we communicate that back to them which, like Nico, this is, gonna be super pertinent to you. In
241 00:21:27.120 ⇒ 00:21:30.290 Patrick Trainer: basically like translating the
242 00:21:30.570 ⇒ 00:21:32.040 Patrick Trainer: development work
243 00:21:32.130 ⇒ 00:21:34.609 Patrick Trainer: and putting that into like
244 00:21:35.510 ⇒ 00:21:41.869 Patrick Trainer: layman’s terms for for our clients, while also providing, like the most transparency as possible
245 00:21:42.443 ⇒ 00:22:00.116 Patrick Trainer: without just like being overwhelming or causing, like any sort of like undue alarm where all while also, like shielding the the business from any sort of like outlash there. So your boots on the ground there
246 00:22:01.275 ⇒ 00:22:09.650 Patrick Trainer: to to put it there. And then the this is more so for our clients in terms of like self service analytics and whatnot.
247 00:22:10.575 ⇒ 00:22:11.460 Patrick Trainer: We’ve
248 00:22:11.510 ⇒ 00:22:19.509 Patrick Trainer: ideally want them to be able to search and explore their data to help them answer their insights faster.
249 00:22:20.090 ⇒ 00:22:28.404 Patrick Trainer: we’re doing a lot of this like with like, I, I think, like real falls into this like hugely
250 00:22:29.150 ⇒ 00:22:41.714 Patrick Trainer: and that also goes into like report generation and scheduling this, I think like this analytics like Number 6 piece, I think, is mostly solved.
251 00:22:42.700 ⇒ 00:22:48.260 Patrick Trainer: but integrating that into the larger picture as a whole. Like I, I still think that there’s like
252 00:22:48.470 ⇒ 00:22:54.910 Patrick Trainer: there’s still pieces missing. There is is kind of what I’m going at. And I think we’re
253 00:22:55.360 ⇒ 00:22:58.869 Patrick Trainer: we like. What does what does that look like in the
254 00:22:59.432 ⇒ 00:23:03.257 Patrick Trainer: like grand scheme of things. It’s like udem that
255 00:23:04.140 ⇒ 00:23:24.200 Patrick Trainer: conversation that we had with Brill the other day of putting clients on like subdomains and then having like their paths. There, that’s yeah. That’s what I’m talking about when when I’m thinking about this in in that like, how does somebody’s real deployment like loop back into brain, forge and relate to that.
256 00:23:24.380 ⇒ 00:23:27.187 Uttam Kumaran: Yeah, like, I think my point on this is
257 00:23:27.990 ⇒ 00:23:34.220 Uttam Kumaran: I want, when whatever we touch this comes to like a design thing also is like, I just want it to be.
258 00:23:35.330 ⇒ 00:23:42.090 Uttam Kumaran: It’s similar to like the iphone is like, you know, that anything on that phone has been thought about. And there’s like a decision behind.
259 00:23:42.120 ⇒ 00:23:52.230 Uttam Kumaran: Right? So I think very similarly. And this goes from everything in the whole company, right from the emails we send to that content to the website to the way we work, it needs to have a consistent
260 00:23:52.470 ⇒ 00:24:08.480 Uttam Kumaran: like feel. And I I think, having this behind Brainforge is great. I mean we’ve already seen clients that we work with, who have don’t care at all, and so who we we wanted to share other stuff through our platform, have a single login, so there are some nice things that we could do there.
261 00:24:08.500 ⇒ 00:24:10.409 Uttam Kumaran: I think the other thing is, some of this
262 00:24:10.420 ⇒ 00:24:22.020 Uttam Kumaran: is going to be client facing some of this is going to be us internally, but I do think there are some business opportunities to to monetize some of the other things, and I think that’ll be that I want to leave
263 00:24:22.040 ⇒ 00:24:37.370 Uttam Kumaran: for later. But it’s 1 of the notes that I written down is, I think there will be an opportunity to actually have this platform be something that is part of like our cell. But also, I think, having stuff like subdomains and things like that is perfect like it makes a lot of sense.
264 00:24:37.370 ⇒ 00:24:38.819 Patrick Trainer: Yeah, exactly.
265 00:24:39.886 ⇒ 00:24:41.180 Patrick Trainer: And then
266 00:24:41.280 ⇒ 00:24:44.930 Patrick Trainer: resources, visualization libraries like, that’s
267 00:24:44.990 ⇒ 00:24:51.593 Patrick Trainer: we’re pretty much like offloading that to to real or providers but that this also could be
268 00:24:52.470 ⇒ 00:25:03.180 Patrick Trainer: like, I mean thinking of it in terms of like white labeling opportunities of that, and just like what you said Udem, with like relating that back to brain forge, and having that be kind of like
269 00:25:03.690 ⇒ 00:25:08.509 Patrick Trainer: that product, and not like real, but like real as a part of
270 00:25:08.760 ⇒ 00:25:12.829 Patrick Trainer: brain forge like that’s kind of where I’m going with this.
271 00:25:13.407 ⇒ 00:25:17.169 Patrick Trainer: And then we are coming into like
272 00:25:17.750 ⇒ 00:25:21.580 Patrick Trainer: ops and keeping all of that stuff alive.
273 00:25:21.650 ⇒ 00:25:25.960 Patrick Trainer: This is also going to encompass like
274 00:25:26.830 ⇒ 00:25:31.189 Patrick Trainer: I mean, I put machine learning, but also thinking about like
275 00:25:31.470 ⇒ 00:25:41.859 Patrick Trainer: AI chat. Gpt, how that dovetails into our operations here. But then this. So a lot of this, like
276 00:25:42.160 ⇒ 00:25:52.650 Patrick Trainer: the 7.1 here is not solved, but partly solved. Like as we’re using like github and actions, and all of that.
277 00:25:53.440 ⇒ 00:25:57.369 Patrick Trainer: So a big piece and big component there we’re we’re we’re good.
278 00:25:57.870 ⇒ 00:26:05.700 Patrick Trainer: But there’s also like when we think about it in the context of all of these other pieces of the the higher platform.
279 00:26:05.820 ⇒ 00:26:08.700 Patrick Trainer: we’re going to have to implement that
280 00:26:08.910 ⇒ 00:26:09.930 Patrick Trainer: as well.
281 00:26:09.990 ⇒ 00:26:12.500 Patrick Trainer: And then monitoring management.
282 00:26:13.010 ⇒ 00:26:21.390 Patrick Trainer: Dbt. Is taking care of a lot of this. But then, again, in the context of of everything else, we need to be aware of that
283 00:26:21.925 ⇒ 00:26:30.320 Patrick Trainer: and understanding how those are going and then I also see this as to as like the brain forge, like site itself.
284 00:26:30.360 ⇒ 00:26:31.730 Patrick Trainer: and
285 00:26:32.590 ⇒ 00:26:39.319 Patrick Trainer: understanding like how that’s working as well as like keeping that in a good
286 00:26:40.680 ⇒ 00:26:44.100 Patrick Trainer: like like best practiced use.
287 00:26:44.760 ⇒ 00:26:48.760 Patrick Trainer: And then we have thoughts around like extensibility.
288 00:26:49.460 ⇒ 00:26:57.040 Patrick Trainer: And so like as time goes on, like, there’s going to be new tools that we’re gonna want to integrate with the larger platform
289 00:26:57.865 ⇒ 00:27:02.620 Patrick Trainer: and so like, how are we integrating those with like
290 00:27:02.820 ⇒ 00:27:08.389 Patrick Trainer: like real and the other tools that we’re using and so
291 00:27:08.890 ⇒ 00:27:20.737 Patrick Trainer: we’re going to need to be like, just cognizant of that and not be, and not like, build ourselves into a box to where, like, we’re impossible to integrate with.
292 00:27:21.140 ⇒ 00:27:24.950 Patrick Trainer: this is, I see, this especially true, too, because, like.
293 00:27:25.100 ⇒ 00:27:30.780 Patrick Trainer: if you think about in terms of our clients like, they’re all they all have their own tools as well.
294 00:27:30.820 ⇒ 00:27:34.399 Patrick Trainer: and we’re going to run into the into the space where
295 00:27:35.215 ⇒ 00:27:39.340 Patrick Trainer: somebody comes. And they’re like, Oh, well, our like, I want
296 00:27:39.690 ⇒ 00:27:46.070 Patrick Trainer: this capability that brain forge is doing. But we we use this platform or like this
297 00:27:46.370 ⇒ 00:27:49.910 Patrick Trainer: Crm or this like visualization library.
298 00:27:50.080 ⇒ 00:27:54.240 Patrick Trainer: we want to use that, but use the other services that brain force does like.
299 00:27:54.450 ⇒ 00:28:01.249 Patrick Trainer: How do we account for that? That’s with these plugins and being able to communicate
300 00:28:01.960 ⇒ 00:28:19.490 Patrick Trainer: with their system and what they already have existing rather than like like, it’s a much harder sell to to get somebody to like. Oh, jettison all the tools that you’re already using and use these like, that’s that’s really difficult. But if we’re like, Oh, yeah, we can just like slip into the
301 00:28:19.540 ⇒ 00:28:23.189 Patrick Trainer: shim into the rest of your tools and
302 00:28:23.280 ⇒ 00:28:29.479 Patrick Trainer: like you won’t even notice then that’s a much, much easier sell. And
303 00:28:29.830 ⇒ 00:28:31.250 Patrick Trainer: there’s the.
304 00:28:31.440 ⇒ 00:28:34.030 Patrick Trainer: this is kind of what I’m thinking about. There.
305 00:28:34.160 ⇒ 00:28:38.429 Patrick Trainer: we have our compliance. And so
306 00:28:39.080 ⇒ 00:28:46.191 Patrick Trainer: that just goes basically into like a lot of like the Rbac schemes as well as
307 00:28:47.280 ⇒ 00:28:48.420 Patrick Trainer: sorry about that
308 00:28:51.990 ⇒ 00:28:55.389 Patrick Trainer: understanding like what’s going on and
309 00:28:56.370 ⇒ 00:29:02.729 Patrick Trainer: setting us up for if we have sock 2 or any sort of it’s basically like to cover your ass
310 00:29:02.760 ⇒ 00:29:08.099 Patrick Trainer: type deal super important, pretty boring. But we all know what it is.
311 00:29:08.130 ⇒ 00:29:11.960 Patrick Trainer: And then we have our knowledge, management
312 00:29:12.150 ⇒ 00:29:13.410 Patrick Trainer: and
313 00:29:13.600 ⇒ 00:29:30.189 Patrick Trainer: collaboration. So like, how are we working together? How are we talking together? And then how are we allowing developers as well as clients to to like, pick up from where someone else left off and be able to like, keep working.
314 00:29:30.530 ⇒ 00:29:30.830 Uttam Kumaran: This.
315 00:29:30.830 ⇒ 00:29:31.140 Patrick Trainer: So let’s.
316 00:29:31.140 ⇒ 00:29:36.194 Uttam Kumaran: This one is is like huge for me. I
317 00:29:36.870 ⇒ 00:29:56.889 Uttam Kumaran: I think we’re gonna continue to get people from a lot of different experiences. But as we grow it’s gonna constantly this is never gonna stop. It’s gonna be a struggle to get people that perfectly fit like the brain forge engineer. So we need to have a process by which.
318 00:29:56.890 ⇒ 00:30:11.000 Uttam Kumaran: in their 1st month, 2 month, 3 months, they can get to that point. You know. So I. So I really wanna push for like having some carve out here for like education, because.
319 00:30:11.490 ⇒ 00:30:15.839 Uttam Kumaran: I think we’re my bias in hiring has never been
320 00:30:15.970 ⇒ 00:30:18.629 Uttam Kumaran: like looking at current
321 00:30:18.710 ⇒ 00:30:23.689 Uttam Kumaran: like skill set. I mainly look for like aptitude and like agency?
322 00:30:24.170 ⇒ 00:30:44.200 Uttam Kumaran: So we’re most likely going to continue to bias towards finding people who can learn versus people who have learned, especially because people who have learned are gonna be really expensive, or they’re already, you know, there’s just not that many of them. So we I want to have a really consistent program by which people come in. And they they just level up to the requirement we need.
323 00:30:44.607 ⇒ 00:30:57.219 Uttam Kumaran: And that doesn’t have to be like, oh, take this like training course, like I think this is, can be something more fluid. And something that actually can take can use client work to kind of make that happen.
324 00:30:57.310 ⇒ 00:31:11.620 Uttam Kumaran: But that’s 1 thing that I think is huge. And then, of course, collaboration and knowledge management. I think we’re gonna have to decide on tooling on, how do we do this? But we have so many docs internally that we do turn around and share with clients
325 00:31:11.620 ⇒ 00:31:29.289 Uttam Kumaran: and content. Repurposing is exactly what we’re doing on the actual content side. And this will be a huge part of that. So any piece of content we produce, internal or external, we want to go through that that sort of process, and I want everybody to write like we are. We are committed to being like a remote
326 00:31:29.510 ⇒ 00:31:37.159 Uttam Kumaran: like Async. 1st company writing is, gonna is our writing or loom or videos or audio is our medium of communication.
327 00:31:37.390 ⇒ 00:31:41.190 Uttam Kumaran: So we have to do that like we just have to push people to, to.
328 00:31:41.710 ⇒ 00:31:53.509 Uttam Kumaran: to prioritize and get used to some of that. And I could tell, like, there’s already some folks who aren’t used to that environment who struggle. So how do we get them? How do we get people to to buy into that? You know, right.
329 00:31:53.510 ⇒ 00:31:54.690 Patrick Trainer: Yeah, exactly.
330 00:31:54.740 ⇒ 00:31:56.880 Patrick Trainer: And so
331 00:31:57.190 ⇒ 00:31:59.640 Patrick Trainer: you’ve basically said everything else
332 00:32:00.170 ⇒ 00:32:03.690 Patrick Trainer: down down here. So we can save on that one.
333 00:32:03.890 ⇒ 00:32:04.980 Patrick Trainer: And so
334 00:32:05.540 ⇒ 00:32:10.535 Patrick Trainer: this is kind of like the high level like strategy. Right? This is,
335 00:32:13.000 ⇒ 00:32:28.319 Patrick Trainer: what there’s like tactics versus this. Yeah, this is strategy, not tactic. And so like the the next piece is like, Okay, like, how do we take this and like distill it into like actual like
336 00:32:28.590 ⇒ 00:32:31.440 Patrick Trainer: work packages and pieces
337 00:32:31.800 ⇒ 00:32:33.240 Patrick Trainer: that’s still
338 00:32:33.290 ⇒ 00:32:38.770 Patrick Trainer: to be determined and is going to be like a lot of the engineering and involved in this.
339 00:32:39.203 ⇒ 00:32:49.020 Patrick Trainer: I do also have like implementation. But then, again, this is just strategy. And it’s not the tactic of like, what are we actually going to do?
340 00:32:49.120 ⇒ 00:32:53.987 Patrick Trainer: And these are like loose guidelines. But I’ve got a hit
341 00:32:54.560 ⇒ 00:32:58.139 Patrick Trainer: loosely roughed out to about 2 the next 2 years.
342 00:32:58.170 ⇒ 00:33:00.100 Patrick Trainer: which I think is like
343 00:33:00.330 ⇒ 00:33:07.451 Patrick Trainer: a reasonable timeline but it also it’s going to depend on like our
344 00:33:08.680 ⇒ 00:33:23.876 Patrick Trainer: I mean, as we scale and scale. Our engineering effort, too, of of how we develop this like a lot of this is like some is simple. Some is not so simple. But I mean it’s like anything. It’s like one piece
345 00:33:24.240 ⇒ 00:33:36.449 Patrick Trainer: of the pie. But this is most I’ve mostly got this laid out in kind of in terms of like the critical path of like what needs to be there first, st before the next thing can
346 00:33:37.090 ⇒ 00:33:39.679 Patrick Trainer: reasonably be thought about being built.
347 00:33:39.900 ⇒ 00:33:41.849 Patrick Trainer: and so like
348 00:33:42.020 ⇒ 00:33:43.730 Patrick Trainer: having our control plane
349 00:33:44.160 ⇒ 00:33:49.020 Patrick Trainer: good to go having catalogs and our rules in our alerting.
350 00:33:49.370 ⇒ 00:33:50.720 Patrick Trainer: That alerting
351 00:33:50.920 ⇒ 00:33:52.849 Patrick Trainer: then leads into
352 00:33:53.850 ⇒ 00:33:57.690 Patrick Trainer: excuse me that quality, monitoring that that pipeline monitoring
353 00:33:59.180 ⇒ 00:34:08.280 Patrick Trainer: And then we have, like more advanced features of like automated run books and our like plugins. And all of that. That’s gonna take while
354 00:34:08.763 ⇒ 00:34:17.959 Patrick Trainer: and then optimization. That’s gonna come last. Just because, like what you build it. And then you and then you optimize it, and so
355 00:34:18.370 ⇒ 00:34:21.747 Patrick Trainer: that’s kind of where where that comes from.
356 00:34:22.409 ⇒ 00:34:30.419 Patrick Trainer: all of this strategy is then hedged back onto like the client roadmap.
357 00:34:38.310 ⇒ 00:34:39.440 Patrick Trainer: pain in the ass.
358 00:34:41.086 ⇒ 00:34:42.239 Patrick Trainer: and so
359 00:34:42.370 ⇒ 00:34:46.580 Patrick Trainer: for new companies like when we’re onboarding them like, what are we
360 00:34:46.620 ⇒ 00:34:52.570 Patrick Trainer: thinking about? And what are we asking ourselves? In understanding their business.
361 00:34:52.630 ⇒ 00:35:01.629 Patrick Trainer: And it’s like, 1st is, understand their business and their objectives and kind of like. And these are more of like the the
362 00:35:01.650 ⇒ 00:35:03.920 Patrick Trainer: tactics that we take in doing that.
363 00:35:04.130 ⇒ 00:35:06.379 Patrick Trainer: understanding what their data is.
364 00:35:06.390 ⇒ 00:35:08.860 Patrick Trainer: mapping to their business objectives
365 00:35:09.570 ⇒ 00:35:11.170 Patrick Trainer: getting their gaps.
366 00:35:11.930 ⇒ 00:35:13.669 Patrick Trainer: How are we collecting it?
367 00:35:13.680 ⇒ 00:35:16.299 Patrick Trainer: How are we going to collect it?
368 00:35:17.740 ⇒ 00:35:19.450 Patrick Trainer: actually doing that
369 00:35:20.350 ⇒ 00:35:27.300 Patrick Trainer: running Dbt and our pipelines, testing everything, documenting everything.
370 00:35:27.350 ⇒ 00:35:36.849 Patrick Trainer: monitoring everything, and then communicating everything. So given that like, this is what we do when onboarding new new clients like.
371 00:35:36.970 ⇒ 00:35:42.819 Patrick Trainer: how do we scale that to multiple clients? Well, that’s where a a platform comes in?
372 00:35:45.720 ⇒ 00:35:50.339 Patrick Trainer: and so then, to diagram all of this out.
373 00:35:52.060 ⇒ 00:35:54.880 Patrick Trainer: they have, like high level components.
374 00:35:55.520 ⇒ 00:36:03.579 Patrick Trainer: might need to just make this. So we have, like our life cycle components over here. So like our data sources ingestion.
375 00:36:03.680 ⇒ 00:36:06.260 Patrick Trainer: consuming those sources
376 00:36:06.280 ⇒ 00:36:08.200 Patrick Trainer: and processing that.
377 00:36:08.750 ⇒ 00:36:15.879 Patrick Trainer: how does all of this fit into our core platform? And that’s like the control plane that I’m talking about.
378 00:36:16.680 ⇒ 00:36:18.799 Patrick Trainer: We’re having those like.
379 00:36:19.680 ⇒ 00:36:28.990 Patrick Trainer: So all of this is pretty much like is going to be like client specific. And so think of this as like then, being like rubber stamped across multiple clients
380 00:36:29.705 ⇒ 00:36:40.879 Patrick Trainer: here like cross cutting concerns. That’s essentially like, what is everything going to have in common with it. So every client is going to need governance, and that compliance
381 00:36:41.203 ⇒ 00:36:47.789 Patrick Trainer: and then monitoring what and alerting. And so all of that is going to gets like packaged and wrapped, backed up
382 00:36:47.820 ⇒ 00:36:52.809 Patrick Trainer: into our core platform and like the the Brainforge platform
383 00:36:53.110 ⇒ 00:36:57.180 Patrick Trainer: going further into it like how our
384 00:36:57.210 ⇒ 00:36:59.730 Patrick Trainer: each of these things broken down.
385 00:36:59.880 ⇒ 00:37:03.240 Patrick Trainer: Let’s see if I can make this even bigger.
386 00:37:03.240 ⇒ 00:37:07.030 Nicolas Sucari: I can see. Yeah, I can see it. Okay, I don’t know. Yeah.
387 00:37:07.564 ⇒ 00:37:17.719 Patrick Trainer: And so, like, we have our. Our big boxes are collaboration and knowledge, our data ingestion, our governance processing.
388 00:37:17.770 ⇒ 00:37:22.159 Patrick Trainer: our consumption. And then we have, like our monitoring and alerting.
389 00:37:22.570 ⇒ 00:37:24.860 Patrick Trainer: Here’s our our warehouse
390 00:37:25.070 ⇒ 00:37:27.010 Patrick Trainer: like Snowflake.
391 00:37:28.370 ⇒ 00:37:35.155 Patrick Trainer: the services that we’re we’re giving so like populating like grill and our bi tools
392 00:37:35.830 ⇒ 00:37:41.460 Patrick Trainer: exposing that all as like an Api to to make that work.
393 00:37:41.730 ⇒ 00:37:43.330 Patrick Trainer: And then
394 00:37:44.240 ⇒ 00:37:56.499 Patrick Trainer: how we ingest different sources data making our catalogs of what’s actually there and then providing quality quality checks as well as lineage which then
395 00:37:57.010 ⇒ 00:37:59.570 Patrick Trainer: takes from our our warehouse.
396 00:37:59.580 ⇒ 00:38:00.274 Patrick Trainer: So
397 00:38:02.860 ⇒ 00:38:05.784 Patrick Trainer: I know that was a lot. But
398 00:38:06.300 ⇒ 00:38:07.800 Patrick Trainer: yeah, thoughts.
399 00:38:09.690 ⇒ 00:38:14.859 Uttam Kumaran: Yeah, I I mean, I think this is like exactly what I was looking for.
400 00:38:16.900 ⇒ 00:38:18.839 Uttam Kumaran: I think, where?
401 00:38:19.260 ⇒ 00:38:26.670 Uttam Kumaran: Like, of course, I think we’re gonna need to think about breaking this up and next steps. But overall like from the business side.
402 00:38:26.720 ⇒ 00:38:29.467 Uttam Kumaran: the the main things I think about
403 00:38:30.350 ⇒ 00:38:32.240 Uttam Kumaran: I think about, how do we
404 00:38:32.770 ⇒ 00:38:35.349 Uttam Kumaran: deliver more for our clients? Faster.
405 00:38:35.420 ⇒ 00:38:58.859 Uttam Kumaran: right? And I think this has a lot to do with that. I do think I do think, like all day about the quality of the folks that we have on the team, especially on the engineering side, and our ability to get engineers and ramp them up and have them work to our level of expectation. And how do we take advantage of stuff that we’ve seen on the edge of, like the most
406 00:38:58.900 ⇒ 00:39:18.596 Uttam Kumaran: you know, sort of stable programs, and bring that to clients who are, who are maybe can’t afford that, or are so small that they would never see something like that. Right? These sorts of things. And I think for context for Nico and I know Roy is listening in on the call as well is these sorts of conversations. And this level of
407 00:39:18.960 ⇒ 00:39:25.549 Uttam Kumaran: sophistication typically only happens at organizations that have had a data platform for, like
408 00:39:25.900 ⇒ 00:39:31.900 Uttam Kumaran: 4 or 5 years, maybe, or longer, and actually have probably like more than
409 00:39:33.230 ⇒ 00:39:38.559 Uttam Kumaran: 20 or 30 or 40 like data. People like these are conversations that
410 00:39:38.670 ⇒ 00:39:52.770 Uttam Kumaran: small companies or companies of data sophistication that we typically work with would never get a chance to take advantage of. So on the business side, like, I am so forward on this because this is what
411 00:39:52.830 ⇒ 00:39:59.400 Uttam Kumaran: ultimately we want to bring to them is stuff that they wouldn’t be able to get in any reasonable timeframe or cost.
412 00:39:59.440 ⇒ 00:40:03.840 Uttam Kumaran: So I’m super excited. Overall. I think the
413 00:40:04.350 ⇒ 00:40:19.829 Uttam Kumaran: for me. I also want to know in terms of budget, not only on people, but on time. And then I also want that to go into like ultimately, like our plan for hiring right?
414 00:40:20.100 ⇒ 00:40:25.109 Uttam Kumaran: I I don’t know. And and this I think we can save for like the recruit recruiting call. But
415 00:40:25.350 ⇒ 00:40:29.320 Uttam Kumaran: I do have like, I do wanna know how we want to
416 00:40:29.360 ⇒ 00:40:38.450 Uttam Kumaran: have engineers like, do we want to have people dedicated to this effort? Do we want to have people that are split in their times?
417 00:40:38.825 ⇒ 00:40:57.919 Uttam Kumaran: You know, I want to understand. Kind of like, what’s the best route before bringing people on. IA lot of this work here. I don’t think that Comp, like the like analytics. Engineers who are just like off the block are gonna be able to wrap their head around. So you’ll need to have some people who are who have worked in a data platform.
418 00:40:58.180 ⇒ 00:41:02.669 Uttam Kumaran: However, I do. I do like the concept of having people
419 00:41:02.720 ⇒ 00:41:07.649 Uttam Kumaran: that are actually engaged with clients also contributing here, because.
420 00:41:07.960 ⇒ 00:41:11.989 Uttam Kumaran: like I don’t know. That seems like really like the way to go
421 00:41:12.550 ⇒ 00:41:17.550 Uttam Kumaran: and, like, you know, dedicating like 10 h or a couple hours a week or a month
422 00:41:18.013 ⇒ 00:41:28.359 Uttam Kumaran: to the platform in addition to to clients, because it keeps things really, really tight. There’s not some brokering information from the client. There’s not a bifurcation of like a client engineer.
423 00:41:28.510 ⇒ 00:41:42.260 Uttam Kumaran: and like a brain like I don’t. That seems like a wrong path for me. And I I don’t. And I think what we’re gonna find is that we’re gonna have external contractors and internal engineers. And maybe that’s where the expectation
424 00:41:42.770 ⇒ 00:41:44.183 Uttam Kumaran: is drawn.
425 00:41:45.080 ⇒ 00:41:50.180 Uttam Kumaran: But yeah, those are all the the thoughts. I think there’s probably a lot more breakdown we need to do. But yeah.
426 00:41:50.180 ⇒ 00:41:50.820 Patrick Trainer: Right.
427 00:41:53.740 ⇒ 00:41:57.549 Nicolas Sucari: And how is like, is there any challenge on
428 00:41:57.570 ⇒ 00:42:01.615 Nicolas Sucari: like if you’re using all of that like our platform
429 00:42:02.330 ⇒ 00:42:08.580 Nicolas Sucari: like, how big has that, and that needs to be in order to manage all of the data from different clients.
430 00:42:08.720 ⇒ 00:42:14.169 Nicolas Sucari: Is that, is there a a challenge there, or is something totally easy to to manage
431 00:42:15.650 ⇒ 00:42:17.530 Nicolas Sucari: more. Turning into the.
432 00:42:17.750 ⇒ 00:42:18.860 Patrick Trainer: How how
433 00:42:18.930 ⇒ 00:42:21.119 Patrick Trainer: easy is this going to be to build.
434 00:42:21.780 ⇒ 00:42:22.420 Nicolas Sucari: Yeah.
435 00:42:24.890 ⇒ 00:42:27.200 Nicolas Sucari: Because it seems like, pretty difficult. Yeah, exactly.
436 00:42:27.200 ⇒ 00:42:29.720 Patrick Trainer: Not. It’s it’s I mean to.
437 00:42:29.720 ⇒ 00:42:35.379 Uttam Kumaran: I wouldn’t say it’s impossible. It’s not. It’s but it’s none of what I’ve read here is not something I haven’t seen before.
438 00:42:35.840 ⇒ 00:42:36.340 Patrick Trainer: Right, right.
439 00:42:36.340 ⇒ 00:42:38.050 Uttam Kumaran: Right like it’s
440 00:42:38.210 ⇒ 00:42:39.120 Uttam Kumaran: so. It’s like.
441 00:42:39.120 ⇒ 00:42:40.820 Patrick Trainer: What I’m detailing is like.
442 00:42:41.170 ⇒ 00:42:43.340 Patrick Trainer: what does a core like
443 00:42:43.380 ⇒ 00:42:52.099 Patrick Trainer: data platform look like? And it’s and it’s this, it’s not a trivial task like this can’t be banged out in a month.
444 00:42:54.410 ⇒ 00:43:00.029 Patrick Trainer: it’s but like, what does a mature system look like? And it’s it’s like this.
445 00:43:00.780 ⇒ 00:43:01.490 Nicolas Sucari: Okay.
446 00:43:01.820 ⇒ 00:43:07.219 Nicolas Sucari: but it it will like, it’s a challenge to build this like, if we are able to build this
447 00:43:07.480 ⇒ 00:43:09.060 Nicolas Sucari: like this is, gonna be awesome.
448 00:43:09.060 ⇒ 00:43:16.160 Uttam Kumaran: Yeah, it’s it’s a not only challenge. Technically, but it’s a challenge or organizationally and like
449 00:43:16.710 ⇒ 00:43:17.720 Uttam Kumaran: and challenge.
450 00:43:17.780 ⇒ 00:43:19.500 Uttam Kumaran: I also
451 00:43:20.040 ⇒ 00:43:26.590 Uttam Kumaran: look, I have talked to a lot of agencies, and frankly, I talk the people who run those are business people.
452 00:43:26.630 ⇒ 00:43:29.880 Uttam Kumaran: and I’ve never spoken with anyone who’s explained
453 00:43:30.220 ⇒ 00:43:31.380 Uttam Kumaran: running
454 00:43:31.450 ⇒ 00:43:35.760 Uttam Kumaran: the like, the internal engineering of an agency like this.
455 00:43:36.930 ⇒ 00:43:41.479 Uttam Kumaran: which I will now be asking more critically of because we have a perspective. But
456 00:43:41.650 ⇒ 00:43:47.419 Uttam Kumaran: I don’t know I don’t see this happening. So the biggest things Nico will have to figure out is.
457 00:43:47.660 ⇒ 00:43:53.030 Uttam Kumaran: at what point in the roadmap do we start seeing either one or more things
458 00:43:53.280 ⇒ 00:43:54.980 Uttam Kumaran: client work, speed up.
459 00:43:56.220 ⇒ 00:43:56.930 Nicolas Sucari: Yeah.
460 00:43:56.930 ⇒ 00:44:21.319 Uttam Kumaran: Costs go down, bugs go down right. We’ll have. Those will be the measurement points. Some of that will be on like our costs, right like, can we deliver more for less? For clients? Some of that will be. We see less bugs on average, some of that will be like it’s what kind of what I mentioned in the chat, which is like? We have client phases, right? Like, what is the 1st month? Look like the 1st 3 months, the 1st 6 months.
461 00:44:21.330 ⇒ 00:44:25.459 Uttam Kumaran: Do we see that cut in half because we’re now on the burn? These are like
462 00:44:25.670 ⇒ 00:44:33.249 Uttam Kumaran: a client who’s on the platform versus not. Those are the the key measures that we’ll be looking at. But I I think already
463 00:44:33.430 ⇒ 00:44:38.840 Uttam Kumaran: we are trying to do some of those things in isolation per client like this just wraps a lot of those up.
464 00:44:40.390 ⇒ 00:44:49.959 Uttam Kumaran: and then it’s again. It’s all quality quality of life stuff, right? Like my ability to go recruit great engineers is gonna be sharing this with like this is like you’re working in.
465 00:44:50.385 ⇒ 00:44:57.110 Uttam Kumaran: My ability to go get great clients is to share that. This is the way that we do typical clients. Right?
466 00:44:57.350 ⇒ 00:45:03.309 Uttam Kumaran: My, you know, my ability to get great salespeople is gonna be sharing that like we, we are confident that we can execute
467 00:45:03.370 ⇒ 00:45:07.390 Uttam Kumaran: on these, no matter if we get 5 or 10 or 15 clients right?
468 00:45:07.630 ⇒ 00:45:08.569 Uttam Kumaran: I think.
469 00:45:09.220 ⇒ 00:45:14.069 Uttam Kumaran: that’s gonna be the biggest this is gonna be kind of like the next step for us.
470 00:45:14.110 ⇒ 00:45:25.230 Uttam Kumaran: I I think we’re gonna have to. This is, gonna be a ramp up like we’re gonna have very limited resources at the start. We’re gonna probably have the same resources we have right now to start on this. But then.
471 00:45:25.420 ⇒ 00:45:32.159 Uttam Kumaran: look, we’re ramping up on everything right? And so as we get more sales, and whether we’re able to measure the investment.
472 00:45:32.320 ⇒ 00:45:34.469 Uttam Kumaran: we can attribute more dollars. So
473 00:45:34.670 ⇒ 00:45:35.440 Uttam Kumaran: yeah.
474 00:45:37.010 ⇒ 00:45:37.610 Nicolas Sucari: Cool.
475 00:45:37.610 ⇒ 00:45:39.669 Uttam Kumaran: It has has to benefit everybody.
476 00:45:39.700 ⇒ 00:45:50.060 Uttam Kumaran: This has to benefit the engineers. It has to benefit. Our our sales efforts. It has to benefit you on you on the actual operations. And like client basically client success side.
477 00:45:50.090 ⇒ 00:45:51.300 Uttam Kumaran: right?
478 00:45:51.470 ⇒ 00:45:57.590 Patrick Trainer: And that’s why that’s why this is all like under the umbrella, under the umbrella of strategy
479 00:45:57.660 ⇒ 00:46:09.229 Patrick Trainer: like, how do we make engineering more efficient. How do we make client onboarding more efficient? How do we make it more resilient? And then all of that
480 00:46:09.460 ⇒ 00:46:23.460 Patrick Trainer: goes back into that sales piece of like. We have an efficient and resilient and compliant system, and so that all creates this like feedback loop that builds into it. And so that’s why
481 00:46:23.670 ⇒ 00:46:27.130 Patrick Trainer: to, I mean, go back to your question, Nico, of like.
482 00:46:27.240 ⇒ 00:46:28.150 Patrick Trainer: how
483 00:46:28.290 ⇒ 00:46:30.479 Patrick Trainer: it’s easy as this is like
484 00:46:30.640 ⇒ 00:46:35.200 Patrick Trainer: it’s it’s not because it’s multifaceted in like
485 00:46:36.290 ⇒ 00:46:38.589 Patrick Trainer: like it’s not. We’re not just
486 00:46:38.790 ⇒ 00:46:43.660 Patrick Trainer: strategizing to solve one thing we’re looking to solve like
487 00:46:43.720 ⇒ 00:46:45.619 Patrick Trainer: 10 different things.
488 00:46:45.690 ⇒ 00:46:56.154 Patrick Trainer: kind of like all wrapped up. But then those 10 things loop back around and influence this. And so it’s like, it’s it’s
489 00:46:56.890 ⇒ 00:47:02.550 Patrick Trainer: There’s a lot of a lot of pieces to it, and a lot of like effectors to it, as well.
490 00:47:06.460 ⇒ 00:47:07.050 Nicolas Sucari: Cool.
491 00:47:07.620 ⇒ 00:47:10.639 Uttam Kumaran: So do we? Wanna I mean, I have
492 00:47:11.670 ⇒ 00:47:14.301 Uttam Kumaran: way too much to think about.
493 00:47:15.220 ⇒ 00:47:17.879 Uttam Kumaran: Do another chat on
494 00:47:18.200 ⇒ 00:47:21.979 Uttam Kumaran: Monday, maybe where we talk about like.
495 00:47:22.300 ⇒ 00:47:28.300 Uttam Kumaran: what do we want to do between now and like maybe the end of the year, and then kind of like, start there.
496 00:47:28.300 ⇒ 00:47:28.880 Patrick Trainer: Right.
497 00:47:28.880 ⇒ 00:47:31.940 Uttam Kumaran: And you know we have.
498 00:47:31.940 ⇒ 00:47:34.169 Patrick Trainer: Is a couple 1 million dollars.
499 00:47:34.170 ⇒ 00:47:34.760 Uttam Kumaran: I you know.
500 00:47:35.250 ⇒ 00:47:35.819 Nicolas Sucari: You’d be.
501 00:47:35.820 ⇒ 00:47:38.559 Uttam Kumaran: Surprise. I think I could go do that for us, but.
502 00:47:39.029 ⇒ 00:47:39.380 Patrick Trainer: Like.
503 00:47:39.380 ⇒ 00:47:41.085 Uttam Kumaran: Life would get a lot more
504 00:47:41.400 ⇒ 00:47:43.230 Uttam Kumaran: hectic around here.
505 00:47:44.640 ⇒ 00:47:57.569 Uttam Kumaran: Oh, I don’t. We’re gonna do with what we can. But at the same time, I do think like this is like this has always been the dream. So maybe let’s talk on Monday and think about like
506 00:47:58.180 ⇒ 00:48:00.650 Uttam Kumaran: how we break down that 1st phase
507 00:48:02.540 ⇒ 00:48:07.540 Uttam Kumaran: And that way, I think, I think explaining this whole thing to folks like Brian and Nick.
508 00:48:07.550 ⇒ 00:48:12.494 Uttam Kumaran: They’re just gonna like it’s gonna fry them. So I wanna talk about what the 1st phase is gonna look like.
509 00:48:12.970 ⇒ 00:48:16.559 Uttam Kumaran: And and then think about okay, like, how do we? Wanna
510 00:48:16.640 ⇒ 00:48:20.909 Uttam Kumaran: how do we want to get some of their time and feedback to start putting resources
511 00:48:20.970 ⇒ 00:48:28.902 Uttam Kumaran: towards this and like actually actioning on it. But I want this to be like a weekly conversation. Basically,
512 00:48:29.470 ⇒ 00:48:33.220 Patrick Trainer: Yeah, yeah, this is definitely not one of those like.
513 00:48:34.020 ⇒ 00:48:35.950 Patrick Trainer: alright I’ve presented. Now, that’s it.
514 00:48:35.950 ⇒ 00:48:36.360 Uttam Kumaran: No, no, no.
515 00:48:36.360 ⇒ 00:48:36.806 Patrick Trainer: And
516 00:48:38.080 ⇒ 00:48:47.430 Uttam Kumaran: Like again. These are all this is like. This whole process has been in both of our heads for a long time, so I’m happy that it’s all it’s all on paper.
517 00:48:47.490 ⇒ 00:48:50.289 Uttam Kumaran: And I do think that
518 00:48:50.766 ⇒ 00:48:56.100 Uttam Kumaran: let’s think about what we can get done with the folks that we have before the end of the year.
519 00:48:56.180 ⇒ 00:49:02.220 Uttam Kumaran: As sales grows, my job will be to get you, you know, more resources and more folks that have it like
520 00:49:02.470 ⇒ 00:49:07.449 Uttam Kumaran: data platform background. The other thing is as we recruit for data engineers.
521 00:49:07.540 ⇒ 00:49:13.259 Uttam Kumaran: pat, like the things we’ll be looking for is some aptitude to contribute to this on the Ae side.
522 00:49:13.660 ⇒ 00:49:18.510 Uttam Kumaran: It’s we’re gonna have less luck there. And on the analyst side, we’re gonna have very little luck.
523 00:49:19.060 ⇒ 00:49:27.729 Patrick Trainer: Like I feel like the ae and analyst side of things like it’s that’s very straightforward and like where the domain starts and ends.
524 00:49:27.730 ⇒ 00:49:33.090 Uttam Kumaran: There may be the they’re gonna the the clients of this or the stakeholders of.
525 00:49:33.090 ⇒ 00:49:34.819 Patrick Trainer: Right? Exactly
526 00:49:34.990 ⇒ 00:49:38.889 Patrick Trainer: exactly and then so it’s it’s like 2 different
527 00:49:39.280 ⇒ 00:49:42.700 Patrick Trainer: frames of mind. It’s like, there’s the
528 00:49:43.670 ⇒ 00:49:49.499 Patrick Trainer: analysts and ae frame of mind of like, okay, how are we modeling client data and getting that done?
529 00:49:49.670 ⇒ 00:49:51.059 Patrick Trainer: But then there’s
530 00:49:51.120 ⇒ 00:49:57.359 Patrick Trainer: this frame of mind of like, okay, how do you enable an analyst and an ae.
531 00:49:57.360 ⇒ 00:49:58.020 Nicolas Sucari: Yeah.
532 00:49:58.020 ⇒ 00:49:59.350 Patrick Trainer: To be able to.
533 00:49:59.390 ⇒ 00:49:59.930 Nicolas Sucari: Paperwork.
534 00:49:59.930 ⇒ 00:50:00.510 Patrick Trainer: Done.
535 00:50:03.330 ⇒ 00:50:05.060 Roy Christian Piñon: Can I? Can! I butt in.
536 00:50:05.320 ⇒ 00:50:06.070 Uttam Kumaran: Yeah.
537 00:50:06.260 ⇒ 00:50:12.408 Roy Christian Piñon: Yeah, appreciate it. So I really appreciate kind of what you’re sharing. And I I was really a
538 00:50:12.840 ⇒ 00:50:18.120 Roy Christian Piñon: I I saw that opportunity. It kind of reminded me as well that Patrick was showing that
539 00:50:18.250 ⇒ 00:50:24.579 Roy Christian Piñon: you always want to go where they start. So I think, kind of art of war here
540 00:50:24.964 ⇒ 00:50:28.970 Roy Christian Piñon: main strategy, I was thinking, was kind of land and expand.
541 00:50:28.990 ⇒ 00:50:34.850 Roy Christian Piñon: I know we got like Patrick mentioned. We have 10 different things we can do for a client.
542 00:50:34.890 ⇒ 00:50:45.449 Roy Christian Piñon: I think value is in the eye of the beholder. So it actually depend on the conversation we’re gonna have, like, what’s really important.
543 00:50:45.610 ⇒ 00:50:48.490 Roy Christian Piñon: What’s the the impact it’s having?
544 00:50:48.560 ⇒ 00:50:55.810 Roy Christian Piñon: We actually start from there. Because because I I don’t see this like rip and replace solution.
545 00:50:55.840 ⇒ 00:50:57.040 Roy Christian Piñon: It’s more of a
546 00:50:57.200 ⇒ 00:51:04.400 Roy Christian Piñon: more of an add on like to have more intelligence on how you’re going to improve your your data management. So
547 00:51:05.020 ⇒ 00:51:12.999 Roy Christian Piñon: I think we can start really small with different clients, because they would typically have their own system.
548 00:51:13.180 ⇒ 00:51:15.789 Roy Christian Piñon: And if we’re introducing something.
549 00:51:16.290 ⇒ 00:51:20.123 Roy Christian Piñon: it’s just where they’re at. Because I was also
550 00:51:21.410 ⇒ 00:51:37.219 Roy Christian Piñon: I heard from with Tom that you mentioned. Like they have. Like these companies, they have different stages of data management, like big companies. They have 30 to 40 people actually managing data. Some of them don’t won’t even have that chance
551 00:51:37.440 ⇒ 00:51:51.949 Roy Christian Piñon: to, you know, utilize the data that they have, and some, you know they might not be that effective. So I guess that’s hitting. We can hit different stages for different companies if we’re going to
552 00:51:51.970 ⇒ 00:52:04.510 Roy Christian Piñon: somehow look at it in that perspective. That hey? What’s kind of the biggest problem you’re having, or could be like more of the attention like, have you even considered analyzing your data.
553 00:52:05.550 ⇒ 00:52:08.340 Roy Christian Piñon: I guess that’s just from how
554 00:52:08.680 ⇒ 00:52:13.240 Roy Christian Piñon: how solutions usually are bought from
555 00:52:13.250 ⇒ 00:52:15.850 Roy Christian Piñon: have, how, how I’ve I’ve
556 00:52:16.270 ⇒ 00:52:18.009 Roy Christian Piñon: I would have experienced it.
557 00:52:18.410 ⇒ 00:52:20.430 Roy Christian Piñon: But let me know.
558 00:52:22.410 ⇒ 00:52:35.639 Patrick Trainer: Yeah, so I think you might have missed the the beginning stages of the of this talk in that like. So this platform is not something that we’re selling to our users. This is a platform
559 00:52:35.680 ⇒ 00:52:43.730 Patrick Trainer: that we’re using internally to service our users. So like the services. Still
560 00:52:44.060 ⇒ 00:53:12.026 Patrick Trainer: like, that’s what we were talking about, like with the the ae and analyst conversation of like that frame of mind versus this higher level frame of mind, and, like the the Ae. And analyst conversation is what exactly what you were saying, like, what are your problems in doing that? But the strategy here of building this internally to service this and then to go back to your like. There’s different companies that different
561 00:53:12.470 ⇒ 00:53:13.890 Patrick Trainer: life stages.
562 00:53:13.940 ⇒ 00:53:15.990 Patrick Trainer: This goes into
563 00:53:16.400 ⇒ 00:53:19.350 Patrick Trainer: like what is Brandforge offering.
564 00:53:19.380 ⇒ 00:53:22.979 Patrick Trainer: And this is the offering of like
565 00:53:23.680 ⇒ 00:53:27.449 Patrick Trainer: for the immature data lifecycle company.
566 00:53:27.700 ⇒ 00:53:31.560 Patrick Trainer: We’re the value that we’re providing is this
567 00:53:32.970 ⇒ 00:53:37.829 Patrick Trainer: very mature system that they then get to use
568 00:53:37.870 ⇒ 00:53:41.833 Patrick Trainer: are are essentially, or then that they then get to
569 00:53:42.750 ⇒ 00:53:46.029 Patrick Trainer: take advantage of. And so
570 00:53:46.330 ⇒ 00:53:47.990 Patrick Trainer: this is all, for
571 00:53:48.430 ⇒ 00:53:55.409 Patrick Trainer: how do we enable our clients to use their data better. Not necessarily like.
572 00:53:55.630 ⇒ 00:53:58.720 Patrick Trainer: is is this something we’re going to implement
573 00:53:58.830 ⇒ 00:54:01.940 Patrick Trainer: multiple times over for each and every
574 00:54:02.100 ⇒ 00:54:06.159 Patrick Trainer: user like that. That’s that’s not the that’s not the case here.
575 00:54:08.940 ⇒ 00:54:09.620 Patrick Trainer: Okay.
576 00:54:09.940 ⇒ 00:54:16.159 Uttam Kumaran: Yeah. So this is actually mostly sorry. My other meeting just got pushed. But this is mostly for our.
577 00:54:16.520 ⇒ 00:54:21.909 Uttam Kumaran: This is like, a mix like it. It benefits more people. But the focus is on
578 00:54:21.940 ⇒ 00:54:29.779 Uttam Kumaran: our internal platform. What what ends up getting surfaced or explained.
579 00:54:30.173 ⇒ 00:54:33.239 Uttam Kumaran: Is up to sales. Right? I think.
580 00:54:33.290 ⇒ 00:54:40.850 Uttam Kumaran: though, but ultimately this will at minimum go to improve the speed and accuracy of our implementations
581 00:54:40.870 ⇒ 00:54:58.589 Uttam Kumaran: at maximum. This could actually be priced in a certain way. Like parts of this can get monetized different ways. But our goal is like we are. We now have. We interact with many clients in different sophistications, and we need a shared platform to interact and speed up and
582 00:54:58.690 ⇒ 00:55:04.150 Uttam Kumaran: not make the same mistakes over and over again. Additionally, there are parts of this system that
583 00:55:04.240 ⇒ 00:55:15.609 Uttam Kumaran: clients that are smaller or medium sized won’t have at their own disposal, and it may take a year or 2 even beyond us to get there. We’re going to give it to them on day one
584 00:55:17.140 ⇒ 00:55:30.719 Uttam Kumaran: and that’s not something that I’ve seen any other sort of analytics. Consulting company offer. Typically, they talk about their process, which is like we come in. We spend a month like diagnosing. We then write models. We then do this?
585 00:55:30.750 ⇒ 00:55:42.697 Uttam Kumaran: There’s nobody that talks in depth about. Oh, actually, from day one you have testing, you have quality metrics. You have a centralized place for governance and security.
586 00:55:43.210 ⇒ 00:55:51.249 Uttam Kumaran: right. And the the reason they don’t talk about those in details because they don’t do that they do what we do so far, which is, we come in and we write models, and
587 00:55:51.360 ⇒ 00:55:53.790 Uttam Kumaran: we do we do what we can.
588 00:55:54.075 ⇒ 00:55:59.650 Uttam Kumaran: This is our ability to come to anybody at at any level and say, like you can, you can scale up.
589 00:55:59.740 ⇒ 00:56:05.539 Uttam Kumaran: You can either come to us and say, like, I want everything, or you could say, we already have our own data governance. We just want you to do
590 00:56:05.580 ⇒ 00:56:09.130 Uttam Kumaran: modeling. But for us internally, we need this to
591 00:56:09.520 ⇒ 00:56:13.640 Uttam Kumaran: to build, you know consistent data models, regardless of the client.
592 00:56:13.640 ⇒ 00:56:16.509 Patrick Trainer: Right. And if you think about it in terms of like
593 00:56:16.700 ⇒ 00:56:22.670 Patrick Trainer: like Sas terms like this is, we’re we’re building a like data platform as a service
594 00:56:23.236 ⇒ 00:56:25.619 Patrick Trainer: essentially. And then and then
595 00:56:26.130 ⇒ 00:56:36.689 Patrick Trainer: going that step further in it, not just being like a a product that you buy. But you’re it’s it’s also looped into a consultancy where we’re
596 00:56:38.230 ⇒ 00:56:42.399 Patrick Trainer: enabling them to actually model the data. And so we’re going a step further.
597 00:56:48.320 ⇒ 00:56:49.050 Roy Christian Piñon: Okay.
598 00:56:54.500 ⇒ 00:56:55.190 Nicolas Sucari: Cool.
599 00:56:57.160 ⇒ 00:57:02.850 Uttam Kumaran: Cool so maybe I think we let’s do a follow up on
600 00:57:03.190 ⇒ 00:57:04.450 Uttam Kumaran: Monday.
601 00:57:04.630 ⇒ 00:57:06.430 Uttam Kumaran: and then we can talk about
602 00:57:07.120 ⇒ 00:57:15.699 Uttam Kumaran: like kind of like 1st steps into this, and I ideally, I want to break down this into like what’s been done.
603 00:57:15.890 ⇒ 00:57:18.439 Uttam Kumaran: what’s been done halfway and like what’s still open.
604 00:57:18.440 ⇒ 00:57:19.030 Patrick Trainer: Right.
605 00:57:19.030 ⇒ 00:57:20.050 Uttam Kumaran: Also, wanna.
606 00:57:20.440 ⇒ 00:57:22.899 Uttam Kumaran: I wanna make some like key
607 00:57:23.060 ⇒ 00:57:25.639 Uttam Kumaran: vendor decisions or have spikes around
608 00:57:28.030 ⇒ 00:57:33.129 Uttam Kumaran: like vendors or tools for specific areas.
609 00:57:34.600 ⇒ 00:57:36.130 Uttam Kumaran: and platforms.
610 00:57:38.850 ⇒ 00:57:39.970 Uttam Kumaran: and then,
611 00:57:40.770 ⇒ 00:57:51.679 Uttam Kumaran: yeah, let’s talk about like what we can accomplish by the end of the year. I think also for Roy, and I’ll I’ll also, you know, kind of probably loop in
612 00:57:52.269 ⇒ 00:57:56.601 Uttam Kumaran: ryan, on the copywriting side is could find some way to turn this into
613 00:57:57.470 ⇒ 00:58:06.450 Uttam Kumaran: You know, conversations with clients and and materials and things for us to to boost. And then in our recruiting meeting next week, we’ll also talk about how we
614 00:58:06.460 ⇒ 00:58:14.277 Uttam Kumaran: can package and discuss this with with people that want to work with us. In particular. Pat, I actually have
615 00:58:15.641 ⇒ 00:58:21.419 Uttam Kumaran: she kind of has worked across the entire stack. But she’s actually more interested in doing analyst work.
616 00:58:21.995 ⇒ 00:58:25.824 Uttam Kumaran: She was a referral from another friend of mine,
617 00:58:26.440 ⇒ 00:58:51.829 Uttam Kumaran: who worked. She worked at spotify, and a couple of the firms, so I would love to introduce you to her. But I was actually kind of waiting for this conversation, because I want actually most of the. I told her that I wanted to introduce her to some of the engineers on the on the call, and I think she’ll be actually very impressed with this. So I think, after the recruiting call next week. I can make that introduction, and that’ll be like the 1st time I think we kind of like try that angle
618 00:58:51.890 ⇒ 00:58:58.810 Uttam Kumaran: again. This is all repurposing of content right? But the one thing I’m proud of is like this is all stuff that I want to talk about like it’s
619 00:58:58.850 ⇒ 00:59:12.509 Uttam Kumaran: this is the stuff that people in our sort of part of the world aren’t talking about and I think a lot of this stuff. Only the most sophisticated companies have any bandwidth to think about this. So I’m really happy that we’re doing that. So.
620 00:59:13.700 ⇒ 00:59:15.120 Patrick Trainer: Yeah, it’s like, this is the
621 00:59:15.160 ⇒ 00:59:18.479 Patrick Trainer: culmination of like 8 years building.
622 00:59:18.480 ⇒ 00:59:19.060 Uttam Kumaran: Yeah.
623 00:59:20.190 ⇒ 00:59:21.929 Uttam Kumaran: no same for me. It’s like, this is like.
624 00:59:21.930 ⇒ 00:59:24.460 Patrick Trainer: Being able to think of it in this way, you know.
625 00:59:24.460 ⇒ 00:59:29.610 Uttam Kumaran: Yeah, and able to see like, you know, after doing this so many times now, not only a brain forge like
626 00:59:29.950 ⇒ 00:59:33.150 Uttam Kumaran: we’ve done this so many times. But even in the past, after
627 00:59:33.200 ⇒ 00:59:34.869 Uttam Kumaran: in a company, you support
628 00:59:35.050 ⇒ 00:59:41.740 Uttam Kumaran: business center after business center after business center, you go to marketing sales everything. So in a given company. I’ve done this so many times that
629 00:59:42.233 ⇒ 00:59:46.929 Uttam Kumaran: you now are able to spot these patterns, and so we’ll build a process around it.
630 00:59:47.000 ⇒ 01:00:02.990 Uttam Kumaran: The people we’re gonna hire, though, are have not are not gonna have experience doing this, that many times. So it’s up to us to kind of build a pro system for them to deliver the same level of quality like, deliver the iphone, no matter if you’re like a junior person or a senior person.
631 01:00:05.150 ⇒ 01:00:12.170 Uttam Kumaran: and so we have a lot of angles to tackle here. But you know I I’m really happy. So I’m excited to kind of put this in motion.
632 01:00:12.780 ⇒ 01:00:13.650 Patrick Trainer: Hell, yeah.
633 01:00:15.700 ⇒ 01:00:16.529 Uttam Kumaran: Cool guys.
634 01:00:16.850 ⇒ 01:00:22.610 Uttam Kumaran: Okay, so I don’t know if we’re I don’t know. Maybe I might talk to you, Nico, later today
635 01:00:23.200 ⇒ 01:00:31.425 Uttam Kumaran: with AI meeting or Pat, if you’re joining that. But Roy, I know it’s late. I don’t know how late you’re staying up, but
636 01:00:32.414 ⇒ 01:00:48.939 Uttam Kumaran: let me know if you wanna join anything else today? And I sent you some stuff on slack. So yeah, and feel free to grab time with Pat and Nico, I wanted I think that basically the across, you know, engineering across operations and across like sales. The 3 of us are really gonna be
637 01:00:49.292 ⇒ 01:01:02.179 Uttam Kumaran: you know, aiming to kind of lead and make decisions around those areas. So I think we’ll have probably another call next week. Mainly on that topic. But I wanted. I want the 4 of us to really be in sync on
638 01:01:02.524 ⇒ 01:01:11.920 Uttam Kumaran: as you can see, there’s many overlaps across all the stuff that we’re working on? But really, how can we continue to make more decisions in each of these areas? So excited.
639 01:01:14.040 ⇒ 01:01:16.450 Uttam Kumaran: cool guys? Thank you.
640 01:01:16.710 ⇒ 01:01:17.310 Patrick Trainer: Later.
641 01:01:18.150 ⇒ 01:01:19.519 Nicolas Sucari: Thanks, guys. Bye-bye.