Meeting Title: Brainforge Final Interview Date: 2026-04-09 Meeting participants: Hussein Diab, Awaish Kumar, Demilade Agboola, Uttam Kumaran
WEBVTT
1 00:00:08.620 ⇒ 00:00:10.230 Awaish Kumar: Hi, Sen, how are you?
2 00:00:10.230 ⇒ 00:00:12.880 Hussein Diab: Oh, hello, doing well. How have you been?
3 00:00:13.370 ⇒ 00:00:14.540 Awaish Kumar: Yeah, I’m good.
4 00:00:15.430 ⇒ 00:00:16.659 Hussein Diab: Good to see you, man.
5 00:00:17.080 ⇒ 00:00:18.269 Awaish Kumar: Good to see you, too.
6 00:00:19.840 ⇒ 00:00:22.630 Awaish Kumar: Yeah, how’s it going so far?
7 00:00:23.960 ⇒ 00:00:30.539 Hussein Diab: Man, it’s, it’s going great. I’m, grateful I’m here. I’m excited to meet the team.
8 00:00:31.090 ⇒ 00:00:35.130 Hussein Diab: Yeah, this, this exercise was fun. I enjoyed it.
9 00:00:37.770 ⇒ 00:00:44.400 Awaish Kumar: So… What did you find, challenging in that exercise.
10 00:00:45.120 ⇒ 00:00:46.090 Hussein Diab: Oh…
11 00:00:50.060 ⇒ 00:00:56.190 Hussein Diab: No big challenges, just, using different, tools,
12 00:00:56.560 ⇒ 00:01:04.330 Hussein Diab: Like, for AirByte, I’ve used the Airflow, I’ve used Fivetran before, I haven’t used AirByte.
13 00:01:04.780 ⇒ 00:01:12.380 Hussein Diab: I wouldn’t say challenging, just different, which, that’s why I enjoyed it, because it’s something new.
14 00:01:13.060 ⇒ 00:01:13.889 Hussein Diab: You know?
15 00:01:14.280 ⇒ 00:01:18.780 Hussein Diab: Github Actions was fun to work with.
16 00:01:19.700 ⇒ 00:01:21.050 Awaish Kumar: Do you have experience?
17 00:01:21.460 ⇒ 00:01:27.289 Awaish Kumar: with, deployments, Of open source tools and things like that.
18 00:01:27.660 ⇒ 00:01:31.699 Hussein Diab: I have experience with open source tools, yes, but…
19 00:01:31.870 ⇒ 00:01:39.609 Hussein Diab: My experience with open source tools, mostly on, like, private clients or private projects.
20 00:01:40.240 ⇒ 00:01:42.599 Awaish Kumar: I mean, in terms of deployments, like.
21 00:01:42.960 ⇒ 00:01:46.970 Awaish Kumar: Airflow, for example. If you use it in your company, you have to deploy it.
22 00:01:47.240 ⇒ 00:01:47.580 Hussein Diab: Yeah.
23 00:01:47.580 ⇒ 00:01:50.290 Awaish Kumar: So you can… everybody in the company can access it.
24 00:01:50.820 ⇒ 00:01:53.300 Awaish Kumar: So, do you have experience with deployments and stuff?
25 00:01:53.870 ⇒ 00:01:59.440 Hussein Diab: No, not with open source, because something like, Airflow
26 00:01:59.640 ⇒ 00:02:03.870 Hussein Diab: we use Fivetran, and it’s not really open source, we…
27 00:02:03.870 ⇒ 00:02:04.560 Awaish Kumar: Pipe?
28 00:02:04.560 ⇒ 00:02:05.140 Hussein Diab: Yeah.
29 00:02:05.140 ⇒ 00:02:10.499 Awaish Kumar: 510 is, like, the… Kind of managed service that is provided to us.
30 00:02:10.699 ⇒ 00:02:11.199 Hussein Diab: That’s right.
31 00:02:11.200 ⇒ 00:02:17.540 Awaish Kumar: I’m all about, like, Asking… yeah, it’s not part of the exercise, I’m just curious, right?
32 00:02:17.790 ⇒ 00:02:18.510 Hussein Diab: Yeah.
33 00:02:18.510 ⇒ 00:02:20.219 Awaish Kumar: Since Airbite was a part of…
34 00:02:20.390 ⇒ 00:02:29.779 Awaish Kumar: exercise. If I were to implement AirByte, I don’t, like, I don’t want to run it on my local machine, right? Obviously, it has to be somewhere,
35 00:02:30.380 ⇒ 00:02:31.330 Hussein Diab: Like a virtual machine.
36 00:02:31.760 ⇒ 00:02:34.160 Awaish Kumar: Posted somewhere so everybody can use it.
37 00:02:34.470 ⇒ 00:02:35.060 Hussein Diab: Yeah.
38 00:02:35.340 ⇒ 00:02:36.090 Hussein Diab: Yeah.
39 00:02:40.540 ⇒ 00:02:41.230 Awaish Kumar: Okay.
40 00:02:42.280 ⇒ 00:02:48.789 Awaish Kumar: Cool, so… Was that… okay, yeah, I drifted from my actual question.
41 00:02:49.150 ⇒ 00:02:53.839 Awaish Kumar: what was, like, the most challenging part? Like, the working with AirByte, or…
42 00:02:54.070 ⇒ 00:03:04.799 Awaish Kumar: I know it’s exciting, but, like, in terms of did you felt any… anything? You had to learn anything, or did you… did it, like, force you to learn about a few things?
43 00:03:08.180 ⇒ 00:03:18.700 Hussein Diab: It’s something I haven’t done before, you know? I’ve done that before, .
44 00:03:22.300 ⇒ 00:03:23.000 Awaish Kumar: Okay.
45 00:03:23.330 ⇒ 00:03:24.030 Hussein Diab: Yo.
46 00:03:24.330 ⇒ 00:03:35.029 Hussein Diab: I… I like it, it’s fun, it’s… it’s working on something else. I, I went an extra step, I created a dashboard as well, just so we can see the data.
47 00:03:35.500 ⇒ 00:03:36.100 Awaish Kumar: Okay.
48 00:03:36.500 ⇒ 00:03:37.100 Awaish Kumar: Boom.
49 00:03:37.100 ⇒ 00:03:38.580 Hussein Diab: Don’t share it with you guys.
50 00:03:38.860 ⇒ 00:03:42.119 Awaish Kumar: Like, how did you create the dashboard? Like, using what tool?
51 00:03:42.520 ⇒ 00:03:47.460 Hussein Diab: I connected the Power BI to my local Postgres.
52 00:03:48.850 ⇒ 00:03:49.420 Awaish Kumar: Okay.
53 00:03:51.270 ⇒ 00:03:58.300 Hussein Diab: So, after DBT, after DBT materialized the view,
54 00:03:58.560 ⇒ 00:04:01.700 Hussein Diab: I tapped Power BI into Postgres.
55 00:04:08.460 ⇒ 00:04:12.569 Hussein Diab: I’ll be waiting for… Put them to join.
56 00:04:13.450 ⇒ 00:04:16.230 Awaish Kumar: Waiting for Utam, but yeah, hi Dami, if you’re…
57 00:04:17.029 ⇒ 00:04:18.500 Awaish Kumar: If you can listen to us.
58 00:04:20.290 ⇒ 00:04:23.150 Demilade Agboola: Hi, I’ll be off-camera today, but I’m here.
59 00:04:25.530 ⇒ 00:04:26.490 Awaish Kumar: Okay.
60 00:04:26.850 ⇒ 00:04:28.900 Demilade Agboola: Alright, Placin, nice to see you again.
61 00:04:29.080 ⇒ 00:04:30.540 Hussein Diab: Nice to see you, Lenny.
62 00:04:30.690 ⇒ 00:04:31.989 Hussein Diab: Or to hear from you.
63 00:04:31.990 ⇒ 00:04:35.149 Demilade Agboola: I was going to say you constantly.
64 00:04:37.130 ⇒ 00:04:42.010 Demilade Agboola: Let’s see if, like, Otam is joining, if not, we can just get started.
65 00:04:46.950 ⇒ 00:04:52.040 Awaish Kumar: Yeah, I think, just wait a… wait for a minute, I…
66 00:05:02.900 ⇒ 00:05:08.420 Awaish Kumar: Yeah, in the meantime, you can start with, so, so, I will just give a brief,
67 00:05:09.670 ⇒ 00:05:15.720 Awaish Kumar: introduction, what this interview will look like. So, we are just…
68 00:05:16.130 ⇒ 00:05:23.690 Awaish Kumar: Going to see the demo, like, what you have worked on, yeah, it will take, like, 5 minutes initially to…
69 00:05:24.010 ⇒ 00:05:32.699 Awaish Kumar: Show us, a brief overview of the RAPO, and how your project is structured, and…
70 00:05:33.170 ⇒ 00:05:39.909 Awaish Kumar: And, if, if, how it looks like, and also, what, if you made any design choices.
71 00:05:40.060 ⇒ 00:05:49.390 Awaish Kumar: Or trade-offs, or anything you want to highlight. And after that, we might have some follow-up Questions on…
72 00:05:50.490 ⇒ 00:06:01.549 Awaish Kumar: on that… specifically on that project, and also the technical questions that arise from that project, like, maybe linked to dbt, or Postgres, and…
73 00:06:01.700 ⇒ 00:06:02.900 Awaish Kumar: And things like that.
74 00:06:03.010 ⇒ 00:06:11.899 Awaish Kumar: Sure. Technical interview, in the end, we might have few, minutes left for you to ask any questions from us, okay?
75 00:06:12.620 ⇒ 00:06:16.600 Hussein Diab: Sounds good. Let me show my screen, we’ll get started.
76 00:06:16.910 ⇒ 00:06:20.519 Awaish Kumar: Yeah, let’s… let’s just start. Like, Utam will join, but yeah, you can start.
77 00:06:21.150 ⇒ 00:06:22.010 Hussein Diab: Sounds good.
78 00:06:26.960 ⇒ 00:06:27.930 Hussein Diab: Alright.
79 00:06:28.910 ⇒ 00:06:41.019 Hussein Diab: Let’s see if here’s something open this one. Cool. Alright, so, I created this on Miro, and this is just steps on what I’ve done to complete the exercise.
80 00:06:41.860 ⇒ 00:06:49.519 Hussein Diab: This is a mermaid flow over here that will tell us the step-by-step on what was done.
81 00:06:49.700 ⇒ 00:06:53.630 Hussein Diab: And this is the steps over here, on what was done.
82 00:06:53.840 ⇒ 00:07:01.970 Hussein Diab: So, let’s say I don’t have Docker installed, or anything installed. The way to get started is to install Docker.
83 00:07:02.090 ⇒ 00:07:09.209 Hussein Diab: And then… Make sure Python is installed, make sure VS Code is installed,
84 00:07:09.420 ⇒ 00:07:14.629 Hussein Diab: That’s my tool of choice. I already have it installed, so I went this route.
85 00:07:15.030 ⇒ 00:07:19.450 Hussein Diab: Make sure Git is connected… And…
86 00:07:19.790 ⇒ 00:07:28.559 Hussein Diab: make sure that I’m able to access the, the branch, and I was able to fork it.
87 00:07:29.150 ⇒ 00:07:31.519 Hussein Diab: And then I created my own.
88 00:07:32.940 ⇒ 00:07:37.369 Hussein Diab: I’m also using PowerShell Terminal.
89 00:07:38.990 ⇒ 00:07:46.330 Hussein Diab: anything like this over here. This is how the tools will look like, and I created my own branch called the Trisane Solution.
90 00:07:47.590 ⇒ 00:08:02.030 Hussein Diab: After environment setup, I made sure my repo is set up, and I cloned the assignment to my own branch, and I accessed this branch in VS Code.
91 00:08:03.060 ⇒ 00:08:15.250 Hussein Diab: One small issue I had. I wrote it down over here, it’s not a big deal, but I was just accessing the parent folder instead of the smaller folder, but I was able to fix it pretty fast.
92 00:08:16.180 ⇒ 00:08:22.330 Hussein Diab: Git status, make sure to tell me that I am in sync with the main branch.
93 00:08:23.010 ⇒ 00:08:26.989 Hussein Diab: And then started… Any questions for me so far?
94 00:08:27.460 ⇒ 00:08:32.159 Awaish Kumar: No, no, okay, I was just, yeah, acknowledging your… what you’re saying.
95 00:08:32.510 ⇒ 00:08:34.140 Hussein Diab: Yeah. Yeah.
96 00:08:34.380 ⇒ 00:08:42.060 Hussein Diab: What else? So, we started with the logic. I downloaded these JSON files.
97 00:08:43.470 ⇒ 00:08:55.179 Hussein Diab: One of the problems, actually, now that you mentioned it, I just remembered I did have one issue, and I have a certain way that I solved it, I will show you. So,
98 00:08:56.120 ⇒ 00:08:58.970 Hussein Diab: Let’s see, Otam’s here. Hello, Otam!
99 00:08:59.540 ⇒ 00:09:06.459 Uttam Kumaran: Hey, so sorry, I just had, like, a client issue that just was laid on. So sorry. Thank you so much for, your patience.
100 00:09:06.850 ⇒ 00:09:12.369 Hussein Diab: purse, I’m going through what I’ve… what I’ve built,
101 00:09:12.370 ⇒ 00:09:13.020 Uttam Kumaran: Okay.
102 00:09:13.020 ⇒ 00:09:16.089 Hussein Diab: for the assessment. So I will continue, and if you have any questions.
103 00:09:16.090 ⇒ 00:09:16.430 Uttam Kumaran: Please.
104 00:09:16.430 ⇒ 00:09:28.650 Hussein Diab: let me know. I was just telling the guys my flow, how I started it, so we went over environment setup, repo setup, breakdown of the challenge.
105 00:09:29.070 ⇒ 00:09:32.079 Hussein Diab: And then for the ingestion layer,
106 00:09:32.290 ⇒ 00:09:38.509 Hussein Diab: We have the JSON file. I downloaded these files, and I created raw schema.
107 00:09:39.120 ⇒ 00:09:45.679 Hussein Diab: Matri F2 set up Airbyte, installed Airbyte, and I created the tables.
108 00:09:45.890 ⇒ 00:09:47.489 Hussein Diab: They’re all tables.
109 00:09:47.880 ⇒ 00:09:53.590 Hussein Diab: One of the issues that I found in AirByte
110 00:09:54.080 ⇒ 00:10:13.109 Hussein Diab: I’m not able to locate… to ingest the files using my own environment, and there is a way to do it, I have to set up the environment in Windows, but I was kind of lazy, so I used the GitHub repo as a source in Airbyte, and I did the sync.
111 00:10:14.410 ⇒ 00:10:26.170 Hussein Diab: Alright. After that, I, assigned the four, Roles for the access control.
112 00:10:26.380 ⇒ 00:10:35.339 Hussein Diab: I set up a BF for Brain Forge developer, one for AirByte, one for dbt, and one for Power BI.
113 00:10:35.650 ⇒ 00:10:44.000 Hussein Diab: I, saved the script in, create rules, SQL under this file, and it’s inside my repo as well.
114 00:10:44.310 ⇒ 00:10:51.549 Hussein Diab: And, I use the least privilege to make sure that the duties are separated. These have…
115 00:10:51.950 ⇒ 00:10:53.680 Hussein Diab: The roles will look like.
116 00:10:56.670 ⇒ 00:10:59.659 Hussein Diab: For the right tree, and all of this.
117 00:11:00.950 ⇒ 00:11:02.599 Hussein Diab: Alright, let’s continue.
118 00:11:03.160 ⇒ 00:11:04.640 Hussein Diab: So,
119 00:11:05.130 ⇒ 00:11:18.019 Hussein Diab: I designed it in a medallion-like architecture. We are using Postgres, so it’s not gonna be called, you know, bronze, silver, gold. So, we have staging, intermediate, and Mart.
120 00:11:18.650 ⇒ 00:11:35.539 Hussein Diab: Using dbt, the files will, land in the row schema, and then, using dbt, we’ll stage them, we’ll take them to intermediate, clean them up, make them gold-ready in March for ingestion.
121 00:11:37.250 ⇒ 00:11:41.020 Hussein Diab: So my final mart is called Order Summary.
122 00:11:41.990 ⇒ 00:11:45.549 Hussein Diab: And my grain is one row per order.
123 00:11:46.100 ⇒ 00:11:48.029 Hussein Diab: This is my columns.
124 00:11:48.340 ⇒ 00:11:56.929 Hussein Diab: I am using dbt plus Postgres, and I will show you my VS Code to go over any of the files.
125 00:11:58.230 ⇒ 00:11:58.820 Awaish Kumar: Okay.
126 00:11:59.400 ⇒ 00:12:08.410 Hussein Diab: Alright, let’s see, what else over here? So after this, I made sure that I can access the data, the data looks good, and then I continued with the actions.
127 00:12:08.870 ⇒ 00:12:18.480 Hussein Diab: So on GitHub, I created a couple of files, one for staging, one for prod, I created a…
128 00:12:18.710 ⇒ 00:12:23.620 Hussein Diab: Oh, I created the secrets inside GitHub, this way it’s not exposed.
129 00:12:24.950 ⇒ 00:12:29.040 Hussein Diab: And, I created the workflow as well.
130 00:12:29.450 ⇒ 00:12:33.650 Hussein Diab: One for prod, or one workflow, that it will take it.
131 00:12:33.750 ⇒ 00:12:37.590 Hussein Diab: This is the secrets here.
132 00:12:37.740 ⇒ 00:12:44.940 Hussein Diab: The secrets will contain 6 variables, user password and database name for staging and for product.
133 00:12:45.920 ⇒ 00:12:49.829 Hussein Diab: And this is my flow, how I created it.
134 00:12:52.010 ⇒ 00:13:08.110 Hussein Diab: And then I validated the data, made sure everything is good, and this is the dashboard that I created. I got a flip sweater here, but… I just, just a way for me to show that the data is correct, that…
135 00:13:08.440 ⇒ 00:13:18.479 Hussein Diab: If we are delivering this to a client, we would be able to tell them what we’re looking for. It looks nicer than just looking at plain data.
136 00:13:18.980 ⇒ 00:13:32.160 Hussein Diab: Some other tools I’ve used, just SQL Workbench to make sure I’m able to access my schema correctly, I’m able to see my database and my tables, and everything looks good.
137 00:13:32.760 ⇒ 00:13:39.930 Hussein Diab: And, what else? This is my VS Code… that I created.
138 00:13:41.450 ⇒ 00:13:48.689 Hussein Diab: And the lineage looks good, too. This is order summary, and lineage will show up correctly from what I used.
139 00:13:51.750 ⇒ 00:13:53.730 Hussein Diab: Yeah?
140 00:13:55.340 ⇒ 00:13:56.760 Hussein Diab: Any questions for me?
141 00:14:00.630 ⇒ 00:14:01.150 Awaish Kumar: Nope.
142 00:14:02.410 ⇒ 00:14:05.749 Uttam Kumaran: Yeah, I think this makes sense. Yeah, I’m interested in… yeah, go ahead, Awish.
143 00:14:06.440 ⇒ 00:14:10.130 Awaish Kumar: Yeah, I was just saying, let’s see the actual project.
144 00:14:10.270 ⇒ 00:14:12.130 Awaish Kumar: the VS Code.
145 00:14:13.440 ⇒ 00:14:16.910 Hussein Diab: Yeah, here’s, here’s the VS Code,
146 00:14:17.560 ⇒ 00:14:25.950 Hussein Diab: This over here, I sent you the link, and I made sure it’s public, so you guys can see the files, and the YAML files, and all of this.
147 00:14:26.310 ⇒ 00:14:34.310 Hussein Diab: Yeah, do you want me to go through each one specifically, how it started?
148 00:14:34.600 ⇒ 00:14:37.620 Awaish Kumar: Nope, I think it’s… Good.
149 00:14:39.660 ⇒ 00:14:42.520 Awaish Kumar: Okay, and, like, how’d you come up with the roles?
150 00:14:42.870 ⇒ 00:14:45.170 Awaish Kumar: Can we talk… talk about those?
151 00:14:45.490 ⇒ 00:14:52.530 Hussein Diab: So… I used the least privilege, I didn’t want to give access to everything, so…
152 00:14:53.020 ⇒ 00:14:57.020 Hussein Diab: I created, this create roles over here.
153 00:14:58.450 ⇒ 00:15:07.600 Hussein Diab: And… similar to any other database, just I have 4 different, roles that I’m using.
154 00:15:07.790 ⇒ 00:15:14.200 Hussein Diab: And I gave specific access. I did have some issues with the AirByte.
155 00:15:14.430 ⇒ 00:15:18.029 Hussein Diab: with a role, and then I fixed it, I added the grant,
156 00:15:19.390 ⇒ 00:15:22.889 Hussein Diab: This one over here, I didn’t add it… or this one, I’m sorry.
157 00:15:24.020 ⇒ 00:15:27.439 Hussein Diab: This one wasn’t added, so I just added it after.
158 00:15:27.590 ⇒ 00:15:29.030 Hussein Diab: To make sure it works.
159 00:15:29.710 ⇒ 00:15:35.600 Awaish Kumar: Okay, and do you think it makes sense to provide, like, login password while creating the role?
160 00:15:41.950 ⇒ 00:15:44.790 Hussein Diab: To create login password while creating the role?
161 00:15:45.060 ⇒ 00:15:45.640 Awaish Kumar: Yep.
162 00:15:45.890 ⇒ 00:15:53.919 Awaish Kumar: So, role is something we don’t want to… want somebody to use the role to log into the Postgres, right? It is just to define.
163 00:15:54.910 ⇒ 00:15:57.889 Awaish Kumar: The different categories of people who can actually
164 00:15:58.170 ⇒ 00:16:04.920 Awaish Kumar: use Postgres database. So, here we are saying that it can log in Choosing a given password.
165 00:16:06.990 ⇒ 00:16:14.339 Hussein Diab: Yeah, maybe for my use case, that’s how I created it.
166 00:16:15.460 ⇒ 00:16:15.800 Awaish Kumar: Okay.
167 00:16:15.800 ⇒ 00:16:17.559 Hussein Diab: We can’t do it without it.
168 00:16:19.420 ⇒ 00:16:24.459 Awaish Kumar: Okay, do you know, like, what we could do if we… if we want to restrict that?
169 00:16:24.980 ⇒ 00:16:28.220 Hussein Diab: Yeah, I would use secrets. I would create secrets.
170 00:16:28.440 ⇒ 00:16:34.569 Hussein Diab: to make sure the roles are not exposed. That’s… that’s how I would usually do it.
171 00:16:34.840 ⇒ 00:16:42.459 Uttam Kumaran: Yeah, I think kind of more of the question is, like, you know, typically, like, we’d work in an environment where you define the role, and then you’re assigning
172 00:16:42.800 ⇒ 00:16:46.550 Uttam Kumaran: you know, the users to a role. Less of, like.
173 00:16:47.590 ⇒ 00:16:54.419 Uttam Kumaran: logging in with the role itself. Is that, like, you know, is that something that you’ve kind of seen a pattern before?
174 00:16:54.650 ⇒ 00:17:03.100 Hussein Diab: Yeah, yeah, I’ve seen it, I use it more in Snowflake, not in PostDeggress. It’s the same idea.
175 00:17:03.360 ⇒ 00:17:09.889 Hussein Diab: That’s how I would usually do it if it was a, like, enterprise kind of project.
176 00:17:10.410 ⇒ 00:17:14.890 Hussein Diab: But for my use case.
177 00:17:16.599 ⇒ 00:17:23.179 Demilade Agboola: I am… I am curious about this, like, would everyone with the role have the same password? Because how would that work?
178 00:17:31.060 ⇒ 00:17:32.279 Hussein Diab: Can you say it again?
179 00:17:34.230 ⇒ 00:17:40.840 Demilade Agboola: So, if we’re assigning a password to a role, would everyone with that role, within that role, have the same password?
180 00:17:40.960 ⇒ 00:17:42.359 Demilade Agboola: And what happens if…
181 00:17:42.960 ⇒ 00:17:48.569 Demilade Agboola: people change roles, does that change passwords? I’m just curious as to how the passwords work with the role.
182 00:17:50.620 ⇒ 00:17:52.870 Hussein Diab: How the passwords work for the world.
183 00:17:59.970 ⇒ 00:18:10.570 Hussein Diab: So the password would work better with the user, not with a role, but…
184 00:18:10.820 ⇒ 00:18:18.890 Hussein Diab: One, I think I was trying to log in somewhere, and it asked me for a password, that’s why I set a password for all of them.
185 00:18:19.190 ⇒ 00:18:21.430 Hussein Diab: But I just made it generic.
186 00:18:22.420 ⇒ 00:18:32.390 Uttam Kumaran: Yeah, I think the way you described it is, you know, typical… that makes… I understood if it’s, like, a… whatever, just for the demo, but yeah, I think mainly on our side, we’re, like.
187 00:18:32.640 ⇒ 00:18:40.549 Uttam Kumaran: Yeah, we set up roles, and then we sort of add users to the roles, and then that way, user can come and go. The role is what inherits all the permissions, right? So…
188 00:18:40.550 ⇒ 00:18:42.870 Hussein Diab: Definitely, definitely, yeah.
189 00:18:43.080 ⇒ 00:18:47.190 Uttam Kumaran: I guess I had a question about, like, materialization strategies, like…
190 00:18:47.410 ⇒ 00:18:56.740 Uttam Kumaran: Wondering if you could walk us through, like, how you materialize some of these tables, and then even just talk about, like, the materialization strategies you’re familiar with.
191 00:18:56.910 ⇒ 00:18:58.760 Uttam Kumaran: What situations you would use what.
192 00:18:59.690 ⇒ 00:19:06.490 Hussein Diab: Yeah, yeah, for sure. So my final materialization for the order summary, it’s creating a view.
193 00:19:06.640 ⇒ 00:19:14.400 Hussein Diab: Let me go to the dbt project, and this will show you over here. So,
194 00:19:14.970 ⇒ 00:19:24.999 Hussein Diab: my staging will be materialized as view, my intermediate will be materialized as view. I set them up in the dbt project.
195 00:19:25.240 ⇒ 00:19:29.690 Hussein Diab: I don’t do this all the time.
196 00:19:29.800 ⇒ 00:19:46.040 Hussein Diab: if it’s a bigger project, and I have a lot of tables in the intermediate, I would create a, like, a config file on top of each dbt SQL code, and I will materialize them from inside.
197 00:19:46.780 ⇒ 00:19:52.169 Hussein Diab: All of these are fuller fresh, like, my tables, it’s a fuller fresh.
198 00:19:52.290 ⇒ 00:20:08.049 Hussein Diab: I have done incremental models before. If we have a big table, we don’t want to truncate it and build it again. My logic over here will start with the
199 00:20:08.940 ⇒ 00:20:15.899 Hussein Diab: with the staging, so I set my sources over here from how they land from AirByte.
200 00:20:16.670 ⇒ 00:20:25.870 Hussein Diab: And then I created, materialization over here for, Customers, orders, and products.
201 00:20:26.470 ⇒ 00:20:31.270 Hussein Diab: And I use the metadata file that you guys have in GitHub.
202 00:20:31.840 ⇒ 00:20:39.299 Hussein Diab: And then… In intermediate, I created order line item and product value.
203 00:20:39.570 ⇒ 00:20:41.799 Hussein Diab: And this is the schema that I use.
204 00:20:43.350 ⇒ 00:20:49.519 Hussein Diab: I also added the test for not null, just to add it, nothing, nothing fancy.
205 00:20:50.020 ⇒ 00:20:51.990 Hussein Diab: I did the same for…
206 00:20:53.880 ⇒ 00:20:58.719 Hussein Diab: staging as well. I added unique and not test, not null and unique.
207 00:21:00.180 ⇒ 00:21:06.290 Hussein Diab: And then for my marts… .
208 00:21:06.290 ⇒ 00:21:08.570 Demilade Agboola: Sorry, sorry to interrupt, I have a question here.
209 00:21:08.670 ⇒ 00:21:18.850 Demilade Agboola: Yeah. If you… as we set up this dbt test, how would you keep track of tests when they fail? Like, how would you be the first to know that this test has failed?
210 00:21:19.570 ⇒ 00:21:32.279 Hussein Diab: So, as of right now, I’m running this in core, so whenever I run it here, I will do dbt build, and this way it will do the seed run and test.
211 00:21:33.550 ⇒ 00:21:36.919 Hussein Diab: This is a demo, so when I do it, I’m gonna get…
212 00:21:36.920 ⇒ 00:21:40.879 Demilade Agboola: In production, like, if you were to deploy this, how would you ensure that you would…
213 00:21:41.080 ⇒ 00:21:42.759 Demilade Agboola: Find out and test for you.
214 00:21:42.760 ⇒ 00:22:00.409 Hussein Diab: I would set user notification, so if I’m using dbt Cloud, for example, then dbt Cloud will set a notification for me, and it will tell me, this column failed because it has, duplicates or something, and then I will fix it
215 00:22:00.620 ⇒ 00:22:24.540 Hussein Diab: I will be the first one to get notified. If we don’t have dbt Cloud, I would use some kind of notification within the environment I’m in. So, for example, if I’m using Azure, I will use ADF to create… to parse out the results, and I’ll send a notification using Logic 7, or anything like this. And same thing for other environments.
216 00:22:25.610 ⇒ 00:22:33.020 Hussein Diab: like, in Databricks, I can use a Python, custom Python code that will send me an email and tell me that this failed.
217 00:22:34.510 ⇒ 00:22:35.710 Demilade Agboola: Okay, thank you.
218 00:22:41.510 ⇒ 00:22:49.870 Uttam Kumaran: Yeah, I’m kind of interested in just, like, in your background, like, what are… have you worked in any, like, really complex sort of dbt environments, and, like.
219 00:22:50.100 ⇒ 00:23:04.140 Uttam Kumaran: anything you can share, whether complexity is in terms of, like, the amount of people collaborating, or in terms of actually, like, the modeling, like, wondering if you could describe any of those. I think it’d be interesting to hear.
220 00:23:04.280 ⇒ 00:23:12.040 Uttam Kumaran: from your background, I just, you know, I was able to read a little bit about your background. I’m sure you worked in some environments with pretty large dbt code bases, or…
221 00:23:12.410 ⇒ 00:23:12.980 Hussein Diab: Definitely.
222 00:23:12.980 ⇒ 00:23:14.110 Uttam Kumaran: physical copies.
223 00:23:14.810 ⇒ 00:23:20.799 Hussein Diab: I want to show you, but I can’t show you.
224 00:23:20.800 ⇒ 00:23:23.189 Uttam Kumaran: Okay, that’s fine, you could describe, that’s okay.
225 00:23:23.190 ⇒ 00:23:32.029 Hussein Diab: Yeah, so Deal News is a company that’s, pretty much fully built on, dbt.
226 00:23:32.470 ⇒ 00:23:36.439 Hussein Diab: they will ingest data using Fivetran.
227 00:23:36.720 ⇒ 00:23:52.180 Hussein Diab: Using many sources, over 100 sources from different datasets, and I’ve built around 250 models there, mixed between views, tables, and incremental models.
228 00:23:53.290 ⇒ 00:24:01.690 Hussein Diab: When I first started on this project, the data was pretty messy, and I cleaned it up.
229 00:24:02.020 ⇒ 00:24:12.279 Hussein Diab: I did a lot of dbt work there, and I set it up to run daily. Our daily run will take, right now 2 hours instead of a really long time.
230 00:24:12.760 ⇒ 00:24:24.570 Hussein Diab: And, yeah, it’s… I’m… I wanna see what else I can tell you about it.
231 00:24:24.570 ⇒ 00:24:30.500 Uttam Kumaran: Yeah, I guess, like, maybe describe to me in abstract, like, what are… what’s the difficulty in working in a, like, a…
232 00:24:30.660 ⇒ 00:24:34.340 Uttam Kumaran: In a system like that, like… Yeah.
233 00:24:34.960 ⇒ 00:24:42.870 Hussein Diab: Well, the difficulty would be managing the DAG, the lineage.
234 00:24:43.280 ⇒ 00:25:00.250 Hussein Diab: just to make sure that if I’m handing over my work to someone else, I will be able to explain the flow. I will be able to explain how the data goes from, ingestion all the way to how it visualized in Tableau.
235 00:25:00.730 ⇒ 00:25:06.059 Hussein Diab: I will have to make sure that I’m writing a lot of comments in my code.
236 00:25:07.490 ⇒ 00:25:12.720 Hussein Diab: I will make sure that, I’m using GET correctly.
237 00:25:12.870 ⇒ 00:25:18.180 Hussein Diab: To sync the data and add the comments on my commits.
238 00:25:18.690 ⇒ 00:25:25.929 Hussein Diab: This way, you know, anyone else want to come after me, they can go to the history, they can… they can see the changes.
239 00:25:25.930 ⇒ 00:25:26.510 Uttam Kumaran: Yeah.
240 00:25:28.150 ⇒ 00:25:34.379 Hussein Diab: The difficulty is really… How complex the data is.
241 00:25:35.290 ⇒ 00:25:41.619 Hussein Diab: You know, the data type, is it structured, is it unstructured?
242 00:25:43.160 ⇒ 00:25:47.199 Hussein Diab: Any of this will add more complexity to it, and of course.
243 00:25:47.500 ⇒ 00:25:55.979 Hussein Diab: The size of the data, and the grain of the data, That’s what makes it complex.
244 00:26:02.160 ⇒ 00:26:04.960 Uttam Kumaran: Yeah, I guess I had a… I was gonna ask another question.
245 00:26:05.450 ⇒ 00:26:09.590 Uttam Kumaran: Sort of about, like, how you gather modeling requirements.
246 00:26:09.870 ⇒ 00:26:14.390 Uttam Kumaran: Like, would love to hear, you know, your opinions on, like.
247 00:26:15.140 ⇒ 00:26:22.519 Uttam Kumaran: you know, we… all of us on the call here have all, you know, done dbt for quite a while. It’s sometimes, like.
248 00:26:22.710 ⇒ 00:26:32.800 Uttam Kumaran: interesting, because, you know, dbt is, like, sometimes you would just, hey, model this mart, you have, like, no idea what’s in something. Sometimes you get a rough scope of a dashboard, and you have to, like, sort of…
249 00:26:33.040 ⇒ 00:26:41.949 Uttam Kumaran: build it towards that, but you still don’t know the specifics. Like, I’m interested if you’ve arrived at any frameworks or thinking about requirements gathering for
250 00:26:42.160 ⇒ 00:26:47.940 Uttam Kumaran: I would say modeling in general, like, dbt is one piece of that, but yeah.
251 00:26:48.280 ⇒ 00:26:55.839 Hussein Diab: Yeah. Definitely. I would have to start with a lot of questions. Sure. Before, before I start on the call.
252 00:26:55.980 ⇒ 00:27:04.299 Hussein Diab: I would research the company, I would research their competitors, I would research their public information.
253 00:27:04.730 ⇒ 00:27:09.280 Hussein Diab: And then, on the call, I will ask them questions regarding their pain point.
254 00:27:09.440 ⇒ 00:27:18.259 Hussein Diab: Because dbt is the tool we use, but the reason we are there is to help them elevate their pain, and to make it easier for them.
255 00:27:18.380 ⇒ 00:27:27.070 Hussein Diab: So I want to make sure that I’m speaking in simple terms, non, you know, technical jargon to make it more harder on them.
256 00:27:27.190 ⇒ 00:27:34.549 Hussein Diab: And then I will ask with why. Why are we here? And then I will ask them what are they comfortable with.
257 00:27:35.830 ⇒ 00:27:43.930 Hussein Diab: The last thing I want to do is to suggest a tool that they are not familiar with, and they don’t have it in their tools.
258 00:27:44.290 ⇒ 00:27:51.820 Hussein Diab: So, if they are on… AWS, then I want to suggest AWS-related tools, and…
259 00:27:52.000 ⇒ 00:28:08.730 Hussein Diab: you know, we’ll go from there. For the modeling, I will ask them about their data grain, their size, how often the candidness of the refresh. This will make me understand at least if we should look at a batch or streaming model.
260 00:28:09.420 ⇒ 00:28:29.399 Hussein Diab: If they are not interested in an hourly-based, canvas, then we don’t need to pay a lot of money and go to the streaming club. We’ll do it batch, we’ll do it daily. Even daily, I would, I would have to ask them, what is this end product serving?
261 00:28:30.660 ⇒ 00:28:38.469 Hussein Diab: Because, you know, believe it or not, we don’t have to do it daily if the data is getting refreshed weekly or monthly.
262 00:28:38.890 ⇒ 00:28:45.960 Hussein Diab: I have a lot of frameworks. One of my favorite framework is the, why, what, to.
263 00:28:46.100 ⇒ 00:28:53.309 Hussein Diab: It’s, I have it on Miro, built on Miro, but I will ask them, why are we here? What are we trying to resolve?
264 00:28:53.540 ⇒ 00:28:56.180 Hussein Diab: Who are the stakeholders?
265 00:28:56.740 ⇒ 00:29:12.810 Hussein Diab: what is our ETL, how it’s currently done, and then questions regarding the Canvas, the files, and all of this. But this will help me zoom out, look at the big picture, and be able to structure it better.
266 00:29:21.860 ⇒ 00:29:24.140 Uttam Kumaran: Any other questions, Awish or Demi?
267 00:29:24.950 ⇒ 00:29:28.380 Awaish Kumar: Yeah, I just wanted to understand, like…
268 00:29:28.880 ⇒ 00:29:33.469 Awaish Kumar: Before this, you mentioned that you have worked on a project which is, like.
269 00:29:35.050 ⇒ 00:29:38.970 Awaish Kumar: Where you have had, like, hmm… More than 250 models.
270 00:29:39.160 ⇒ 00:29:46.830 Awaish Kumar: So, my question is, regarding how… Were you monitoring the… data quality?
271 00:29:48.960 ⇒ 00:29:52.319 Awaish Kumar: And also, in terms of data pipeline, like, how…
272 00:29:52.450 ⇒ 00:30:01.390 Awaish Kumar: You are monitoring both the things that the… The models are running Making sure they are…
273 00:30:01.990 ⇒ 00:30:07.550 Awaish Kumar: running on time, and they are working fine. Second thing is monitoring of data quality.
274 00:30:07.760 ⇒ 00:30:16.180 Awaish Kumar: That data freshness, And the other, like, business-related questions… Yeah.
275 00:30:16.640 ⇒ 00:30:32.600 Hussein Diab: So, data freshness is one of the most important things we do, but we don’t want to add it to all of the data sources. So, I added freshness on specific seed files, specific sources that we’re getting to.
276 00:30:32.610 ⇒ 00:30:48.600 Hussein Diab: And we will get a warning if the data is not fresh every single day, and then we will get a notification if it took longer. This will help us decide if we should ignore it, or we should continue forward.
277 00:30:49.650 ⇒ 00:30:55.640 Awaish Kumar: Okay, so… How that flow looks like, how you get the notification, where… Right.
278 00:30:56.010 ⇒ 00:31:10.469 Hussein Diab: So, the… we’re using dbt Cloud, so the notification will be set up using an email. So, I will get an email, and it will tell me this data is not fresh. This… this data source is not fresh.
279 00:31:11.260 ⇒ 00:31:11.850 Awaish Kumar: Okay.
280 00:31:12.090 ⇒ 00:31:17.870 Hussein Diab: For the quality, I am using tests, different, type of…
281 00:31:18.000 ⇒ 00:31:21.049 Hussein Diab: dbt test that will tell me
282 00:31:21.150 ⇒ 00:31:25.670 Hussein Diab: what I’m looking for, and a lot of time, for example, if we’re expecting an email.
283 00:31:26.050 ⇒ 00:31:31.380 Hussein Diab: And we get that data there, it will break the whole pipeline, so…
284 00:31:31.790 ⇒ 00:31:39.370 Hussein Diab: We’ll have to go fix it, and the more we do it, the more sturdy and better our pipeline will be in the long term.
285 00:31:39.970 ⇒ 00:31:48.289 Awaish Kumar: Okay. Yeah, since we have 250 models, and then a lot of quality checks, it’s possible that you get a lot of emails.
286 00:31:48.420 ⇒ 00:31:55.660 Awaish Kumar: So, what is your process for triaging that… those issues, and… And closing… close… closing those.
287 00:31:57.680 ⇒ 00:32:06.769 Hussein Diab: So we don’t get a lot of emails, because these 250 models weren’t all built at once. They were built in sequence.
288 00:32:07.440 ⇒ 00:32:15.010 Hussein Diab: So we will get an email about a specific model that will break the pipeline downstream.
289 00:32:15.220 ⇒ 00:32:17.470 Hussein Diab: And then we’ll go and fix it.
290 00:32:19.290 ⇒ 00:32:24.100 Awaish Kumar: I mean, if I have a, for example, a check on a revenue column.
291 00:32:24.350 ⇒ 00:32:27.259 Awaish Kumar: If it detects an outlier, send me a refund.
292 00:32:28.570 ⇒ 00:32:29.830 Awaish Kumar: a notification.
293 00:32:30.090 ⇒ 00:32:31.980 Awaish Kumar: Yeah. What happened in those cases?
294 00:32:33.240 ⇒ 00:32:46.640 Hussein Diab: So I would go and I would investigate. I would look at the data upstream. I would see if the data is coming from the source bad, or it’s something in MySQL code that was not done correctly.
295 00:32:47.780 ⇒ 00:32:52.620 Hussein Diab: But we have this… like,
296 00:32:53.160 ⇒ 00:33:01.740 Hussein Diab: Rails on the way to make sure that the data is, you know, landing correctly in the end of the pipeline downstream.
297 00:33:03.940 ⇒ 00:33:09.819 Hussein Diab: And, most of these models were incremental.
298 00:33:09.960 ⇒ 00:33:13.690 Hussein Diab: So, a folder refresh is only done
299 00:33:13.940 ⇒ 00:33:20.140 Hussein Diab: If we ever need it, but full refresh will take around 18 hours to complete.
300 00:33:23.650 ⇒ 00:33:24.350 Awaish Kumar: Okay.
301 00:33:29.480 ⇒ 00:33:34.030 Awaish Kumar: Okay, then how would you test it, a model like that? Like, for example.
302 00:33:35.420 ⇒ 00:33:41.670 Awaish Kumar: We have a model which takes 18 hours to execute, and… Hmm.
303 00:33:42.510 ⇒ 00:33:49.270 Awaish Kumar: And there was… there is a… some kind of change, like, where we are… Business definition, metric definition changed.
304 00:33:49.500 ⇒ 00:33:51.590 Awaish Kumar: And you have to redone it.
305 00:33:52.030 ⇒ 00:33:52.480 Hussein Diab: Okay.
306 00:33:53.920 ⇒ 00:34:05.670 Awaish Kumar: And also, like, how would you run it? Like, normally we do go from, like, development environment, then staging, then production, so how will you handle this kind of scenario where it takes, like.
307 00:34:06.190 ⇒ 00:34:09.480 Awaish Kumar: So, like, 18 hours to actually run it.
308 00:34:09.480 ⇒ 00:34:16.659 Hussein Diab: So, one thing that I recommend when we do a full refresh is to go and up our cluster size.
309 00:34:16.790 ⇒ 00:34:22.740 Hussein Diab: I don’t want us to use the same cluster that we’re on daily. So, we’ll increase our cluster.
310 00:34:23.239 ⇒ 00:34:34.839 Hussein Diab: increase the resources that we have in Redshift, and then I will run a full refresh, and instead of 18 hours, it will take 2 hours. That’s one of the ways that I would increase it.
311 00:34:35.159 ⇒ 00:34:40.430 Hussein Diab: Yeah, but I wouldn’t want to run it in chunks.
312 00:34:40.690 ⇒ 00:34:52.979 Hussein Diab: At all, because the data is not going to line up. So if I run a full refresh on certain models on Monday, and then I go Tuesday and I run a full refresh there.
313 00:34:53.350 ⇒ 00:34:56.639 Hussein Diab: my data is not gonna match, so I wanna make sure that…
314 00:34:57.240 ⇒ 00:34:57.580 Awaish Kumar: Young.
315 00:34:57.580 ⇒ 00:34:58.240 Hussein Diab: I have it all.
316 00:34:58.240 ⇒ 00:34:58.820 Awaish Kumar: Cheers.
317 00:34:59.050 ⇒ 00:35:04.879 Awaish Kumar: My question was more like, how would you test it? Like, would you just go run in production directly, or…
318 00:35:05.010 ⇒ 00:35:07.570 Awaish Kumar: Will you run it with staging and then production?
319 00:35:07.570 ⇒ 00:35:24.110 Hussein Diab: No, we have a dev environment, and then we have abroad, so I would test it in dev. I can also test it in my own branch on my local PC if needed, but what I usually do, I test it in dev, make sure everything is good to go, and then I will push my branch.
320 00:35:24.280 ⇒ 00:35:34.339 Hussein Diab: I will have someone else also review it, even if I’m the expert, even if I’m doing everything. This is one of the habits that I like to,
321 00:35:34.640 ⇒ 00:35:40.520 Hussein Diab: add on the team is for at least one other person to look at your code, validated.
322 00:35:41.030 ⇒ 00:35:50.220 Hussein Diab: And of course, we do have a CICD, and once it passes, it will… I will merge it, and then I will run a full refresh and prompt.
323 00:35:51.320 ⇒ 00:35:51.960 Awaish Kumar: Okay.
324 00:35:59.390 ⇒ 00:36:07.890 Demilade Agboola: My question was just gonna be about… What’s your experience like handling
325 00:36:08.030 ⇒ 00:36:13.030 Demilade Agboola: Cold data, like, data that you’ve had to… not use.
326 00:36:13.580 ⇒ 00:36:18.079 Demilade Agboola: frequently? Yeah. What architecture do you store it in, and how do you…
327 00:36:18.440 ⇒ 00:36:26.109 Demilade Agboola: go about the process of using… like, when it’s cold, how do you get it? Like, when you need it regularly, how do you get it from cold storage?
328 00:36:26.330 ⇒ 00:36:31.210 Demilade Agboola: to, your warehouse, where it’s more actively used, and…
329 00:36:32.710 ⇒ 00:36:38.560 Demilade Agboola: like, basically, how do you think about that process? What infrastructure do you use, and how do you go about that?
330 00:36:39.450 ⇒ 00:36:49.569 Hussein Diab: So, if I’m on BigQuery, for example, I will use cold versus hot. If something I’m using more often, then I will store it in a different database.
331 00:36:49.690 ⇒ 00:36:56.969 Hussein Diab: And… with the clients that I have, Everything is hot.
332 00:36:57.490 ⇒ 00:37:01.350 Hussein Diab: We haven’t… we don’t… we don’t do anything cold.
333 00:37:01.900 ⇒ 00:37:07.060 Hussein Diab: For the cold stuff, we use dbt seed. We use them as a seed file.
334 00:37:07.370 ⇒ 00:37:12.039 Hussein Diab: But for architecture purposes, this is, this is,
335 00:37:12.210 ⇒ 00:37:18.810 Hussein Diab: what I would do. I would store it in, like, a near line or a far line, which is, like, cold or hot.
336 00:37:20.100 ⇒ 00:37:23.849 Hussein Diab: This is one of the questions that I got on my GCP exam.
337 00:37:26.880 ⇒ 00:37:33.349 Hussein Diab: It’s… this exam is great, but a lot of the use cases, I don’t use them day to day, but…
338 00:37:33.350 ⇒ 00:37:48.900 Uttam Kumaran: No, totally. I mean, so… but it’s actually, like… I think about this in Snowflake a lot, that’s my… I’ve done Snowflake for a long time, and you just need it. When you need it, though, it’s, like, very good that you’re like, I think… I remember this, like, one parameter. It’s like…
339 00:37:49.050 ⇒ 00:38:01.520 Uttam Kumaran: Otherwise, you’re like, you’re gonna Google, and you’re gonna get, like, the wrong thing, because there’s even stuff today where I’m like, oh yeah, like, in 2019, I remember this thing I did, like, let’s try that.
340 00:38:02.910 ⇒ 00:38:04.740 Hussein Diab: happens later. Definitely.
341 00:38:06.840 ⇒ 00:38:22.179 Hussein Diab: Most recently, in my type of work, I’ve been doing more, like, leading the clients, asking them the right questions, asking them about, their pain, and also doing a lot of automation.
342 00:38:22.490 ⇒ 00:38:28.389 Hussein Diab: So, if… They’re using… they’re doing something, and…
343 00:38:29.270 ⇒ 00:38:39.920 Hussein Diab: it’s saving them a long time to get it done. This is where I will step in, I will tell them, hey, there is an easier way to implement this, and that will save you, I don’t know, like…
344 00:38:40.640 ⇒ 00:38:48.609 Hussein Diab: 6 hours a week that all of this can be done automatically, so you don’t have to have a headache and deal with it.
345 00:38:48.830 ⇒ 00:38:53.280 Hussein Diab: That’s what I’ve been more, helping my clients with.
346 00:39:00.620 ⇒ 00:39:05.810 Uttam Kumaran: Any other questions, Awish or Demi? I was gonna maybe allow… yeah, yeah, go ahead.
347 00:39:06.640 ⇒ 00:39:12.120 Awaish Kumar: Yeah, I was saying the same. We don’t have any questions, so we are, like, almost on time.
348 00:39:12.230 ⇒ 00:39:15.639 Awaish Kumar: So, yeah, if you have any questions for us, sign.
349 00:39:16.460 ⇒ 00:39:18.349 Hussein Diab: Yeah, what do you think of the dashboard?
350 00:39:19.820 ⇒ 00:39:21.000 Awaish Kumar: Yeah, it looks great.
351 00:39:23.070 ⇒ 00:39:24.860 Uttam Kumaran: I like it. How did you, like,
352 00:39:25.110 ⇒ 00:39:27.810 Uttam Kumaran: Put it together, or, like, what was the process?
353 00:39:29.070 ⇒ 00:39:31.670 Hussein Diab: First I created the logo.
354 00:39:32.120 ⇒ 00:39:32.650 Uttam Kumaran: Nice.
355 00:39:32.650 ⇒ 00:39:38.920 Hussein Diab: I asked ChatGPT to create a logo for me, and I made it, seem like.
356 00:39:38.920 ⇒ 00:39:39.490 Uttam Kumaran: Nice, yeah.
357 00:39:39.490 ⇒ 00:39:40.649 Hussein Diab: out of Brain Forge.
358 00:39:40.650 ⇒ 00:39:42.720 Uttam Kumaran: Yes, I saw that.
359 00:39:43.290 ⇒ 00:39:47.530 Hussein Diab: Chatgbt also created my hex colors.
360 00:39:47.910 ⇒ 00:39:48.360 Uttam Kumaran: Nice.
361 00:39:48.360 ⇒ 00:39:51.449 Hussein Diab: So it will look good altogether.
362 00:39:51.660 ⇒ 00:40:08.800 Hussein Diab: My process, I created containers, and I put the containers all together, and I do this a lot with my clients, so, like, I do have experience on what they’re looking for, in terms of filters and colors and, you know, stuff like that, but…
363 00:40:08.950 ⇒ 00:40:11.480 Hussein Diab: Yeah.
364 00:40:11.480 ⇒ 00:40:22.040 Uttam Kumaran: Well, tell me about, like, I guess, you know, one of the questions I always love asking candidates is, like, what’s… what’s sort of still exciting you about, like, data work? Like, what’s next?
365 00:40:22.290 ⇒ 00:40:35.559 Uttam Kumaran: like, where are you trying to get deeper on? Where do you feel like, okay, I’ve always wanted to sort of get this, but never got a chance to? Like, one of… when folks come into Brainford, I always am like.
366 00:40:35.750 ⇒ 00:40:44.609 Uttam Kumaran: what is giving you energy, and then where can I continue to challenge, you know? So, even just your reflections on that would be really, really helpful.
367 00:40:44.910 ⇒ 00:40:53.929 Hussein Diab: Yeah, my most favorite part of this work is having an impact on the human being.
368 00:40:54.220 ⇒ 00:41:01.949 Hussein Diab: So, data is data, tools are tools, but what this dashboard can serve in the end of the day.
369 00:41:02.540 ⇒ 00:41:14.999 Hussein Diab: is what really excites me. When the person comes in, and they will look at the refund rate, and they will say, alright, I have 35% refund rate, I’m doing something wrong, let me fix it.
370 00:41:17.100 ⇒ 00:41:22.379 Hussein Diab: Seeing this slide pops up on the people, that’s what really excites me.
371 00:41:22.560 ⇒ 00:41:32.500 Hussein Diab: Data-wise, I love… exploring something new I haven’t used before, so…
372 00:41:32.720 ⇒ 00:41:37.849 Hussein Diab: And I enjoy this type of work, so just for fun, I do stuff that…
373 00:41:38.150 ⇒ 00:41:42.850 Hussein Diab: excites me. So, for example, last weekend, I was working on…
374 00:41:43.010 ⇒ 00:41:56.619 Hussein Diab: having my data from Fitbit, from my watch, and stream it using GCP, and have a streaming data that will create a Looker,
375 00:41:58.190 ⇒ 00:41:58.869 Uttam Kumaran: Yeah, yeah, yeah.
376 00:41:58.870 ⇒ 00:42:09.190 Hussein Diab: You know, I haven’t used Looker before, so I was like, I just want to use it to use it so I can learn how to use it, and if someone asks me, I’ll say that, yeah, I’ve used Looker before.
377 00:42:09.190 ⇒ 00:42:10.439 Uttam Kumaran: Yeah, yeah, yeah.
378 00:42:10.440 ⇒ 00:42:13.600 Hussein Diab: It’s not my favorite, but I know how to use it.
379 00:42:13.900 ⇒ 00:42:14.460 Uttam Kumaran: Yeah.
380 00:42:14.640 ⇒ 00:42:19.350 Hussein Diab: So, yeah, this, this always excites me, and
381 00:42:19.560 ⇒ 00:42:34.549 Hussein Diab: what also excites me is how we can leverage AI to do even more. I’m working on, like, an experiment with MCPs and how we can implement them. I’ve used a lot of…
382 00:42:34.880 ⇒ 00:42:37.520 Hussein Diab: DAGs…
383 00:42:38.140 ⇒ 00:42:51.419 Hussein Diab: like, something like, Notebook LM, or, you know, I experimented with the clients, I showed them how we can use it, how we can leverage it, and I’ve got a lot of really good responses there.
384 00:42:51.880 ⇒ 00:42:58.470 Hussein Diab: And also, what excites me, actually, the most is helping people do this.
385 00:42:58.880 ⇒ 00:43:06.499 Hussein Diab: So… Mentoring and teaching Whether it’s clients or someone internally, like, hey.
386 00:43:06.630 ⇒ 00:43:14.360 Hussein Diab: Let’s do it together, let’s work on it, let me show you what good looks like, so we can build something better together.
387 00:43:14.670 ⇒ 00:43:15.230 Uttam Kumaran: Yeah.
388 00:43:15.420 ⇒ 00:43:16.080 Hussein Diab: Yeah.
389 00:43:17.500 ⇒ 00:43:33.419 Uttam Kumaran: Tell me, like, what… what has been interesting about meeting our team? And I’m sorry, I know I should give you some time to answer… ask some questions, but what was… what’s been interesting about meeting our team, or what initially drew you to Brainforge, or what have you learned through the process? And I think, naturally.
390 00:43:33.710 ⇒ 00:43:45.859 Uttam Kumaran: I’m happy… we’re more than happy… this is my last meeting today, so I’m more than happy to stay on as long as you’d like to answer any questions, but I’m just curious, like, how the process has been, and what’s been interesting, what’s been cool.
391 00:43:46.100 ⇒ 00:43:46.720 Hussein Diab: Yo.
392 00:43:47.150 ⇒ 00:43:52.800 Hussein Diab: First of all, Kayla did a great job reaching out to me.
393 00:43:53.280 ⇒ 00:43:55.469 Hussein Diab: She introduced me to the company.
394 00:43:55.600 ⇒ 00:44:09.299 Hussein Diab: And, I honestly wasn’t planning into moving or doing anything like that, but what interested me the most is when Aesh told me how you guys serve small to medium clients.
395 00:44:09.460 ⇒ 00:44:19.250 Hussein Diab: And this is really interesting for me, because at my current role, we serve Fortune 500 companies, and it’s huge companies that…
396 00:44:20.160 ⇒ 00:44:30.889 Hussein Diab: They have a lot of money, they have a lot of work to do, but having an impact on these, like, moms and pops shops, or mid-sized companies.
397 00:44:31.130 ⇒ 00:44:31.700 Uttam Kumaran: Yeah.
398 00:44:31.920 ⇒ 00:44:35.350 Hussein Diab: I think the work that we do there
399 00:44:35.490 ⇒ 00:44:38.309 Hussein Diab: We’ll have much larger impact on that.
400 00:44:38.720 ⇒ 00:44:39.410 Hussein Diab: So…
401 00:44:39.410 ⇒ 00:44:41.020 Uttam Kumaran: It doesn’t get lost, for sure.
402 00:44:41.270 ⇒ 00:44:45.880 Hussein Diab: Yeah, that’s what really draws me to a brain fort.
403 00:44:46.700 ⇒ 00:44:51.480 Hussein Diab: That was my first question to Arish, like, hey, who are your clients? Who do you serve?
404 00:44:52.020 ⇒ 00:44:52.530 Uttam Kumaran: Sure.
405 00:44:52.730 ⇒ 00:44:55.590 Hussein Diab: What else?
406 00:45:03.410 ⇒ 00:45:06.850 Hussein Diab: I want to see what else throws me to the theme.
407 00:45:08.000 ⇒ 00:45:13.480 Uttam Kumaran: It’s fine, it’s just… I know we haven’t spent a lot of time together, so… but even hearing that is really good.
408 00:45:13.480 ⇒ 00:45:13.920 Hussein Diab: Yeah.
409 00:45:13.950 ⇒ 00:45:32.920 Uttam Kumaran: I… I… we’re, I mean, it means a lot hearing that, like, Awash and Kayla and Demi and everybody on our team means a lot to us, and we are a people business, like, ultimately, we monetize our expertise, but our clients buy from us, not… we’re not like a dev shop.
410 00:45:33.050 ⇒ 00:45:42.129 Uttam Kumaran: these are… we’re not, like, subcontractors. These are all owned relationships, like, our… we will have relationships with the CEO or the C-suite.
411 00:45:42.160 ⇒ 00:45:57.399 Uttam Kumaran: And we’re not a big company, and we are able to actually go displace a lot of other companies… a lot of consultancies that are way bigger than us, because we’re really, really good, technically. We’re really fast, but we’re partners for these businesses. We’re not just, like.
412 00:45:57.850 ⇒ 00:46:14.829 Uttam Kumaran: here’s 10, here’s 100 tickets, blah blah blah. We don’t take… we don’t do that type of work. And so it means a lot for us to bring on clients where we… we know, we’re confident we can make an impact, and that we also know that they value the work that we’re doing.
413 00:46:14.990 ⇒ 00:46:19.550 Uttam Kumaran: You know, my background is in engineering. We’ve always been confident in our engineering.
414 00:46:19.630 ⇒ 00:46:31.769 Uttam Kumaran: prowess. I think more of what we’re learning about is, like, and Demi and Awash and I, like, we do the 30th, the 40th DBT implementation. Okay, how do we make that next experience better? Like, what did we learn?
415 00:46:31.770 ⇒ 00:46:41.560 Uttam Kumaran: So the next customer that walks through the door, like, has a better, you know, they’re ramping up, they’re understanding of, like, where data modeling fits into the entire journey.
416 00:46:41.710 ⇒ 00:46:46.519 Uttam Kumaran: And then, to your point, we are super, super, super AI-forward.
417 00:46:46.730 ⇒ 00:47:00.569 Uttam Kumaran: In that we are using it not only to improve the speed, the quality, like, we are doing things for smaller and mid-sized businesses that it’s only typically the enterprise, has, because they just have
418 00:47:00.720 ⇒ 00:47:07.759 Uttam Kumaran: tons of people, tons of time. We are bringing that to… to everybody. We have really deep partnerships with
419 00:47:08.010 ⇒ 00:47:11.589 Uttam Kumaran: Snowflake, with a lot of the vendors in the space, because
420 00:47:11.840 ⇒ 00:47:23.740 Uttam Kumaran: we sell their tools better than they do, like, we’re… we know everything about… about them, right? And so, I don’t know, I feel like we… we’ve… we’ve really tried to… to… to focus on that a lot, so… but it’s… for us, it’s like…
421 00:47:23.890 ⇒ 00:47:30.439 Uttam Kumaran: people are still buying from people, so, you know, I think it really matters, especially as consultants, you know.
422 00:47:31.450 ⇒ 00:47:33.010 Hussein Diab: Yeah, I love that.
423 00:47:33.310 ⇒ 00:47:52.030 Hussein Diab: I think I’ve seen one of your videos or description, and it’s saying how, you care about the people, you care about the growth, and, making them better, and this is… this is really important for me,
424 00:47:52.240 ⇒ 00:47:58.700 Hussein Diab: when I joined the call, whether with Demi or Awash, I actually shared with them this,
425 00:47:59.080 ⇒ 00:48:08.280 Hussein Diab: This over here, And, it’s a way to start the conversation by, hey, I’m a person, it happens.
426 00:48:08.280 ⇒ 00:48:08.730 Uttam Kumaran: Yes, hello.
427 00:48:08.730 ⇒ 00:48:13.109 Hussein Diab: that I know how to use data, and I know how to use tools, but…
428 00:48:13.260 ⇒ 00:48:18.640 Hussein Diab: This is me. That’s what I will be talking to you about when we’re on a call.
429 00:48:18.760 ⇒ 00:48:22.060 Hussein Diab: Yeah.
430 00:48:23.640 ⇒ 00:48:25.340 Uttam Kumaran: I’m here in Austin, too, by the way.
431 00:48:25.560 ⇒ 00:48:28.870 Hussein Diab: I know, I know, what’s your favorite place in Austin?
432 00:48:29.420 ⇒ 00:48:30.559 Uttam Kumaran: In terms of food?
433 00:48:30.560 ⇒ 00:48:31.300 Hussein Diab: Yeah.
434 00:48:31.560 ⇒ 00:48:35.089 Uttam Kumaran: Dude, that’s, like, a tough… that’s, like, not an easy click.
435 00:48:35.090 ⇒ 00:48:35.780 Hussein Diab: It’s not.
436 00:48:35.780 ⇒ 00:48:38.030 Uttam Kumaran: You gotta say, top 3, like…
437 00:48:38.800 ⇒ 00:48:42.629 Uttam Kumaran: Like, you gotta give me, like, a genre or something, like, and that’s too…
438 00:48:42.630 ⇒ 00:48:45.870 Hussein Diab: That question’s too hard. What’s your favorite fajita, please?
439 00:48:46.660 ⇒ 00:48:51.939 Uttam Kumaran: Bajita Place, okay, there’s some good Tex-Mex. There’s a place called Matt’s El Rancho, Tex-Mex.
440 00:48:52.220 ⇒ 00:48:54.650 Uttam Kumaran: There’s a couple of other good Tex-Mex.
441 00:48:55.210 ⇒ 00:48:58.720 Uttam Kumaran: I don’t know if there’s any good Lebanese food…
442 00:48:59.140 ⇒ 00:49:01.479 Uttam Kumaran: Here, but there is some good, like.
443 00:49:02.030 ⇒ 00:49:07.239 Uttam Kumaran: there is some good shawarma, there is some good… there’s a Palestinian,
444 00:49:07.490 ⇒ 00:49:12.660 Uttam Kumaran: lunch buffet place that I go to, because they have baklava is so much baklava.
445 00:49:13.660 ⇒ 00:49:19.079 Uttam Kumaran: Yeah, but no, that’s good. I agree, like, I think… For us.
446 00:49:19.160 ⇒ 00:49:26.040 Uttam Kumaran: you know, I think a lot about the few people in my career that invested in me, and took time.
447 00:49:26.090 ⇒ 00:49:43.510 Uttam Kumaran: to show me the way to do things versus just you do it and come show up to the meeting. And ultimately, I also think about all the people that didn’t do that. And so at our company, I think really, really hard about how to not be those people. Like, and again, I think…
448 00:49:43.670 ⇒ 00:49:58.339 Uttam Kumaran: I… I’m still… I still do a lot of client work, and I think about, okay, how do I set the team up for success? And technology and things change, yeah, and we’re… but we’re all smart, and so how can I make sure everybody comes alongside us?
449 00:49:58.430 ⇒ 00:50:07.489 Uttam Kumaran: You’ll also see that we’re a consultancy with very little in terms of formal, like, project management, account management, like, all of the engineers here are, like.
450 00:50:07.690 ⇒ 00:50:09.620 Uttam Kumaran: Even more full stack.
451 00:50:09.720 ⇒ 00:50:29.270 Uttam Kumaran: Like, we all manage linear, we all manage client-facing documents, we do everything, and I also felt like that as an engineer, you always are, like, get put into the corner, like, just go write the code, don’t come… don’t do the business side. That’s not, like, what the situation is, at all. Like, this is an engineer first class.
452 00:50:29.820 ⇒ 00:50:48.949 Uttam Kumaran: citizen situation, and that’s, like, I take a lot of responsibility for that, and it’s difficult, like… but it takes all of us learning about the commercial side, learning about the people side, you know, how to… and for us, we’re delivering outcomes. Yes, DBT, Snowflake, BI, all of that, but
453 00:50:49.200 ⇒ 00:51:07.069 Uttam Kumaran: we’re delivering, like, that someone can accomplish their OKRs, you know, in the right time frame, company can grow, you know, they can save some money, like, so we start with the outcomes, and then we have a host of different bells and whistles to help them achieve that. But yeah, I mean, it’s…
454 00:51:07.270 ⇒ 00:51:14.420 Uttam Kumaran: or business is, like, only people, like, there’s, like, yeah, I think that otherwise you could go to a hundred other places, right? So…
455 00:51:15.220 ⇒ 00:51:16.140 Hussein Diab: Love that.
456 00:51:16.260 ⇒ 00:51:20.469 Hussein Diab: Love that. Otam, how do you get your clients?
457 00:51:21.720 ⇒ 00:51:41.379 Uttam Kumaran: Yeah, it’s a great question. So, it’s changed a lot, since, you know, we were… we were sort of small. Now we have several different, you know, core sales channels. I would say, like, 30-40% of our business comes through partnerships, whether that’s leads that are being, you know, sort of co-sold with us.
458 00:51:41.590 ⇒ 00:51:52.800 Uttam Kumaran: We do a lot of co-marketing with partners, so there’s a huge partnerships piece of our business. That’s something that’s, like, really close to me, because as an engineer, I was always very opinionated about the tools and the vendors we use.
459 00:51:53.040 ⇒ 00:51:59.709 Uttam Kumaran: And so naturally, we only implement the stuff that we love using, if we can.
460 00:51:59.860 ⇒ 00:52:08.260 Uttam Kumaran: And that would… naturally, we love using a few tools, and one day I called them, said, we love implementing you, I’m bringing you guys a lot of money, like, we should collaborate.
461 00:52:08.310 ⇒ 00:52:22.690 Uttam Kumaran: So there’s partnerships. We also do a lot of, like, just direct selling. So, we have a lot of experience in e-commerce, in software, in a few other industries. We have case studies, so we can go and
462 00:52:22.800 ⇒ 00:52:27.790 Uttam Kumaran: Generally, you know, are able to get meetings and say, how can we be helpful?
463 00:52:28.050 ⇒ 00:52:40.219 Uttam Kumaran: And then we’re also now, because, you know, we started to do a lot in AI in the last few years, we’re starting to win business, helping people with workflow automation, building AI agents.
464 00:52:40.650 ⇒ 00:52:42.460 Uttam Kumaran: And so in that way.
465 00:52:42.510 ⇒ 00:52:51.569 Uttam Kumaran: we’re able to use things internally, and then start to commercialize them externally, which is really, really nice. But it’s a lot of credibility building.
466 00:52:51.600 ⇒ 00:52:55.890 Uttam Kumaran: It’s a lot of, like, really long-term thinking, like.
467 00:52:55.900 ⇒ 00:53:12.969 Uttam Kumaran: there are folks who I will talk to every few months for… for years, until we… somehow there’s an opportunity, and you’ll see that some people… someone we’ve talked to, like, a year and a half ago will be like, hey, I just switched companies, I thought about you guys, I would love to hire you.
468 00:53:13.460 ⇒ 00:53:15.799 Uttam Kumaran: Is that sales? Is that just…
469 00:53:15.820 ⇒ 00:53:27.199 Uttam Kumaran: being helpful, you know, it’s a relationship, and so we try to be helpful first. We try to be very non-transactional. This is a business.
470 00:53:27.200 ⇒ 00:53:44.779 Uttam Kumaran: But I think we really are, like, confident in the work that we do. And so, really, we’re… we go to clients and we say, like, how can we be helpful? Like, what’s painful? Ideally, we’re not… and we don’t typically do work with people that are like, cool, we want to fire everybody, we need data to do it. It’s all growth.
471 00:53:44.900 ⇒ 00:53:48.799 Uttam Kumaran: related challenges, typically. So, how do we go more efficiently?
472 00:53:48.920 ⇒ 00:53:55.709 Uttam Kumaran: Our business is half a million, we want to get to a billion. Data, we don’t have a great data foundation.
473 00:53:55.760 ⇒ 00:54:11.789 Uttam Kumaran: We want to migrate to this… to Snowflake in order to improve our data foundation. Those are the kind of stories. And then the last piece is, like, I think we’re unique in that we’re very full-stack data. Everything from data warehousing to modeling, ETL,
474 00:54:11.970 ⇒ 00:54:13.010 Uttam Kumaran: to…
475 00:54:13.230 ⇒ 00:54:27.400 Uttam Kumaran: the BI tools to product analytics to strategy, like McKinsey-level, you know, actual revenue and decision strategy, is what we do. And I think it’s rare to go to a company where you have, like, all of that breadth, and it’s not, like.
476 00:54:27.630 ⇒ 00:54:34.280 Uttam Kumaran: we do any of those, like, poorly in one thing better. Like, I think we’re… we’re pretty good, and we’re really tight, so…
477 00:54:34.620 ⇒ 00:54:48.939 Uttam Kumaran: you know, there’s a lot of challenges sometimes internally in data teams that do those handoffs, right? From the warehousing and the landing to modeling, to then the BI, to then finally there’s an analyst who’s 5 steps away from, like, why a metric was…
478 00:54:49.270 ⇒ 00:54:54.110 Uttam Kumaran: Because it’s just us internally, you can jump… you can make that leap really fast with us.
479 00:54:54.220 ⇒ 00:55:13.220 Uttam Kumaran: And, you know, now I’m just telling you kind of our pitch, but, like, that’s sort of what we tell people. And it’s… collectively, in our company, we’ve done this for a long time, and we’re really confident that if you give us a shot, like, we will, in the first, you know, as fast as we can, we will prove… prove our value.
480 00:55:13.330 ⇒ 00:55:19.040 Uttam Kumaran: And that’s just sort of, like, how… how we do things. So we have… we’re really, really sticky.
481 00:55:19.130 ⇒ 00:55:36.979 Uttam Kumaran: I think most of our clients tend to expand with us over time, unless some unforeseen, you know, thing. And we’re a young company, too. We’re only, like, two and a half years old, and I… I think, like, even in terms of, like, having any significant team size, even less… less than that, right? So…
482 00:55:37.200 ⇒ 00:55:45.959 Uttam Kumaran: one thing I would say is, like, it’s a testament to our people, and how aggressive we’ve been to try to, like, aim to…
483 00:55:46.440 ⇒ 00:56:05.559 Uttam Kumaran: compete with the biggest companies is what you’re seeing. Like, I never looked at ourselves similar to, like, other small agencies or freelancers. That’s not who we’ve ever been. I’m thinking about how do we compete with Deloitte or Accenture. I only think about that. I don’t think about the other people in our size, like.
484 00:56:05.820 ⇒ 00:56:15.570 Uttam Kumaran: That’s… those guys are great, fine. I’m like, how can… how can you… how can a client be like… I talked to Deloitte, but then I talked to the Brainforge guys, and they seem better, like…
485 00:56:15.940 ⇒ 00:56:29.719 Uttam Kumaran: that’s… that’s… that’s the level that I would like to be at. And so, we’ve… I’ve had that since it was just me sitting on a laptop, and I don’t know, you can ask Awaysh and Demi have both been here for over a year, like, I feel like it’s…
486 00:56:29.890 ⇒ 00:56:36.679 Uttam Kumaran: it’s not… I don’t know if I’ve changed, but I don’t know, Awish or Demi, if you guys have any thoughts on that as well, and, like, what…
487 00:56:36.780 ⇒ 00:56:41.089 Uttam Kumaran: what customers tend to like about us, or anything you want to share with Usain?
488 00:56:43.310 ⇒ 00:56:47.849 Awaish Kumar: Yeah, I think it has been, like, personally a great experience for me, being here.
489 00:56:48.020 ⇒ 00:56:50.960 Awaish Kumar: I’ve been here for a year now, and
490 00:56:51.300 ⇒ 00:56:55.710 Awaish Kumar: And the client that we work with is… are kind of, like, really like what we do.
491 00:56:55.920 ⇒ 00:57:04.259 Awaish Kumar: I’m… Yeah, we have… like, clients from different industries, maybe, healthcare, CPG,
492 00:57:04.690 ⇒ 00:57:12.300 Awaish Kumar: But we are able to successfully apply our expertise in different domains and for different size of clients.
493 00:57:12.670 ⇒ 00:57:16.030 Awaish Kumar: And I had great feedback from all of them.
494 00:57:18.570 ⇒ 00:57:27.520 Demilade Agboola: Yeah, I agree. I think I definitely agree with everything said so far. We try our best to provide quality as well as speed.
495 00:57:27.860 ⇒ 00:57:33.650 Demilade Agboola: Because we want the client to be able to get value for whatever money they’re paying.
496 00:57:33.860 ⇒ 00:57:37.200 Demilade Agboola: And to be able to get, results.
497 00:57:38.300 ⇒ 00:57:44.239 Demilade Agboola: Because that’s why they came here anyway, because they were having issues. So the faster they can see results, the better they are able to…
498 00:57:44.340 ⇒ 00:57:46.019 Demilade Agboola: Find value for that money.
499 00:57:46.380 ⇒ 00:57:49.350 Demilade Agboola: Yeah, I… yeah, definitely.
500 00:57:49.670 ⇒ 00:57:57.719 Demilade Agboola: It’s always important for us to be as in their faces as possible, and even internally, we try our best to
501 00:57:58.170 ⇒ 00:58:01.609 Demilade Agboola: Have processes that allow us to show
502 00:58:01.780 ⇒ 00:58:09.039 Demilade Agboola: The impact of what we’re doing, because the idea of, like, for instance, building out entire data pipeline with a dashboard.
503 00:58:09.210 ⇒ 00:58:11.809 Demilade Agboola: If no one uses it, and we just, like.
504 00:58:12.100 ⇒ 00:58:16.089 Demilade Agboola: Build it. There was no point in all of that, process.
505 00:58:16.090 ⇒ 00:58:19.770 Uttam Kumaran: Which is pretty common that, you know, we’ve all been part of teams where it’s like.
506 00:58:20.570 ⇒ 00:58:24.780 Uttam Kumaran: What was this… what was the point of this, you know? Yeah.
507 00:58:25.110 ⇒ 00:58:28.440 Hussein Diab: But if we do a good job, then, you know, they will tell us.
508 00:58:28.440 ⇒ 00:58:33.569 Uttam Kumaran: The whole thing, even if you’re at the 1-yard line, you have to touch… you have to score.
509 00:58:33.980 ⇒ 00:58:41.230 Uttam Kumaran: So that’s the thing, and we know that about data, because it’s this whole manufacturing pipeline, but if you mess up the dashboard.
510 00:58:41.390 ⇒ 00:58:45.820 Uttam Kumaran: Doesn’t matter. Nothing matters. But, it’s like, it’s like I…
511 00:58:49.710 ⇒ 00:58:50.889 Demilade Agboola: Let’s on me, you’re muted.
512 00:58:54.470 ⇒ 00:58:57.569 Uttam Kumaran: Yeah, so that’s… that’s the main thing I wanted to cover, yeah.
513 00:58:58.830 ⇒ 00:59:10.940 Hussein Diab: Oh, thank you guys, I appreciate the response. I have, like, 15 questions listed over here. I was gonna ask you more and more about it. I don’t know how much time do you have?
514 00:59:11.410 ⇒ 00:59:13.629 Hussein Diab: How is your utilization rate?
515 00:59:18.390 ⇒ 00:59:20.979 Uttam Kumaran: Sorry, I was just switching my microphone. Can you say it one more time?
516 00:59:21.280 ⇒ 00:59:24.919 Hussein Diab: Yes, I was asking about your utilization rate.
517 00:59:25.310 ⇒ 00:59:36.089 Uttam Kumaran: Yeah, I would say we’re over, like, we’re… this is why we’re hiring. I think we have plenty of work coming in right now. I think what we’re doing internally, is we are…
518 00:59:36.360 ⇒ 00:59:49.310 Uttam Kumaran: one thing I always tell the team is, like, and you can even… you can ask Sueish and Demi, like, I think we’ve all played many different roles since we started at Brainforge, and I am always a big proponent of
519 00:59:49.620 ⇒ 00:59:59.709 Uttam Kumaran: promoting with… from within. And so right now, we’re going kind of through a transformation where we’re promoting and creating a leadership class in our delivery system.
520 01:00:00.250 ⇒ 01:00:01.320 Uttam Kumaran: And…
521 01:00:01.690 ⇒ 01:00:20.579 Uttam Kumaran: right now, I feel like my job is to bring on both people that can take on the daily work and ladder up to these leaders, but also people that can lead. And so, yeah, I mean, I think anywhere… we typically sit anywhere from, like, 80% to 100% utilization. I think we’re anywhere
522 01:00:20.790 ⇒ 01:00:28.560 Uttam Kumaran: our goal is to sit, kind of, between 80 and 90. And then the way I think about it, and again, this is where, like, I don’t come from a formal consulting background, but I’m…
523 01:00:29.110 ⇒ 01:00:40.979 Uttam Kumaran: very, very aware of all the metrics that are in this world, and I think for us, because of how we use AI, because of the way we deliver our service and how we price.
524 01:00:41.200 ⇒ 01:00:48.400 Uttam Kumaran: We can be a lot more flexible in focusing on delivering outcomes, not, like.
525 01:00:48.820 ⇒ 01:00:51.980 Uttam Kumaran: Maxing hours every week, and not, like…
526 01:00:52.400 ⇒ 01:00:59.810 Uttam Kumaran: you know, I don’t know if there’s a single time where I’ve been like, oh, the client didn’t pay for this, so we’re not gonna do it. It’s like…
527 01:01:00.000 ⇒ 01:01:04.170 Uttam Kumaran: I don’t think, like, I don’t know, it’s, like, not even in my DNA to say stuff like that, so…
528 01:01:04.420 ⇒ 01:01:11.929 Uttam Kumaran: we… we… but now… but again, as an operator, I’m very aware of utilization in the business, and so…
529 01:01:12.000 ⇒ 01:01:30.910 Uttam Kumaran: Yeah, I would say anywhere from 80% to 90%. I think these guys have been with us during this kind of period of growth, so it’s been a little bit crazy. But that’s probably where we’ll land. And in the remaining time, like, we… we’re doing a lot of internal L&D training, you know, we’re gonna… we’ll start to invest more in the team.
530 01:01:30.970 ⇒ 01:01:42.789 Uttam Kumaran: And so there’ll be time that goes to just, like, upskilling, training, you know, culture building between pods and things like that. Which, again, is, like, I don’t expect us… like, I’m not interested in…
531 01:01:42.950 ⇒ 01:02:02.049 Uttam Kumaran: saying, like, okay, we’re gonna go hire hundreds of people. Like, I’m interested in making the biggest impact where we keep the bar high. So if we can’t keep the bar high, we won’t… we just won’t grow. We won’t grow headcount. So this is where, from my background in software, I don’t think about headcount in the way I think in consulting, it’s very common.
532 01:02:02.170 ⇒ 01:02:05.939 Uttam Kumaran: I think there has to be a really non-linear relationship
533 01:02:06.080 ⇒ 01:02:07.740 Uttam Kumaran: And in fact, I think
534 01:02:07.800 ⇒ 01:02:23.290 Uttam Kumaran: you know, I could see in the folks on this call, we are able to do a lot more than maybe your JD or your pre-pass roles may define you as, and I think we’re our… Brainforge is a team of those types of people, where we’re… most people are at least
535 01:02:23.350 ⇒ 01:02:35.619 Uttam Kumaran: pretty good in two of the domains, if not 3, 4, 5. Like, everybody on this call has done most of the data, full-stack data, you know? So, I’m less interested in, like.
536 01:02:35.630 ⇒ 01:02:44.610 Uttam Kumaran: okay, we… every client needs a project manager, every… we need 10 people per client, like, I don’t have rules like that. I… I’m more… I’m interested in client success.
537 01:02:44.670 ⇒ 01:02:53.859 Uttam Kumaran: I’m interested in having tight teams where you build trust, and we run async. We’re a remote, you know, we’re a remote-first company.
538 01:02:53.960 ⇒ 01:03:10.029 Uttam Kumaran: And we’re focusing on outcomes, yeah. But it’s hard, it’s a culture. I’ve been part of a lot of remote teams and a lot of engineering teams, and if your broader company isn’t like that, you may feel like your team is a certain way, but the company is, like, not allowing you to grow. I think
539 01:03:10.130 ⇒ 01:03:20.320 Uttam Kumaran: I feel like I love working here. It’s a lot of super, super smart people, and I try to… I always try to think about, would I join Brainforge?
540 01:03:20.400 ⇒ 01:03:39.409 Uttam Kumaran: You know, and that’s sort of my… my barometer. If I was on the outside, would I interview here, and would I join, and would I be happy? And so I feel like I’m trying every day to make that a reality. And so your question was about utilization. That’s a long-winded answer by saying, I feel like we’re pretty standard in terms of utilization.
541 01:03:39.410 ⇒ 01:03:41.360 Hussein Diab: I appreciate the response.
542 01:03:41.360 ⇒ 01:03:41.890 Uttam Kumaran: Yeah.
543 01:03:41.890 ⇒ 01:03:48.229 Hussein Diab: Yeah, I’m gonna stop here with my questions. I think that’s a.
544 01:03:48.230 ⇒ 01:03:53.100 Uttam Kumaran: That’s fine, if you have other… I’m happy, I don’t know the other guys… I’m happy to answer questions, yeah, anytime.
545 01:03:53.100 ⇒ 01:03:59.499 Hussein Diab: I’ll do one more, I’ll do one more. Okay, please. And this is important to me, that’s why, that’s why I want to ask it.
546 01:03:59.500 ⇒ 01:04:00.150 Uttam Kumaran: Sure, sure, sure.
547 01:04:00.150 ⇒ 01:04:02.189 Hussein Diab: And you kind of…
548 01:04:02.650 ⇒ 01:04:09.880 Hussein Diab: said it in a way, but I would love to hear from you about the mission and the vision.
549 01:04:10.190 ⇒ 01:04:11.200 Hussein Diab: Why?
550 01:04:11.370 ⇒ 01:04:14.730 Hussein Diab: Did you… did you create Brainforge? Why does it exist?
551 01:04:15.590 ⇒ 01:04:18.830 Uttam Kumaran: Sure. Awish, I know if you have to drop, feel free.
552 01:04:19.020 ⇒ 01:04:19.560 Awaish Kumar: Okay.
553 01:04:20.020 ⇒ 01:04:22.130 Uttam Kumaran: Okay, thanks, dude.
554 01:04:22.720 ⇒ 01:04:36.690 Uttam Kumaran: Yeah, it’s a great question. It’s not often I get asked. I usually am, like, I’m sharing. But I’ve worked in engineering, you know, most of my career. I started my career as a BI engineer.
555 01:04:36.780 ⇒ 01:04:47.110 Uttam Kumaran: Wework, right when that company was scaling, a lot, so I had a lot of experience early on. I then went to a company that was doing QR codes.
556 01:04:47.320 ⇒ 01:04:51.360 Uttam Kumaran: basically the month before COVID, and then that business sort of really grew, and I was…
557 01:04:51.590 ⇒ 01:05:09.450 Uttam Kumaran: the only data person. I built the whole data team. And then, you know, did some data product work there, and then, towards the next job, I actually led product at a data company, data transformation company. It was a zero to one. I was, like, one of the first employees, built the first product.
558 01:05:09.700 ⇒ 01:05:15.509 Uttam Kumaran: So I love data. I feel like it’s my calling in terms of…
559 01:05:15.650 ⇒ 01:05:20.739 Uttam Kumaran: technical and the business side, and I really love that it’s, like.
560 01:05:21.110 ⇒ 01:05:27.690 Uttam Kumaran: really, ultimately, we’re helping people make more decisions and more accurate decisions, and I really believe that.
561 01:05:27.800 ⇒ 01:05:35.349 Uttam Kumaran: But I was sort of left with a choice. I worked in startups, like, my whole career, and when I left that job.
562 01:05:35.410 ⇒ 01:05:55.080 Uttam Kumaran: I was like, what do I do next? I… I had done some brief stints in consulting or, like, kind of contracting. Never really impressed with what I saw when I went into that world. I was good, and so, like, you can kind of hold on to your own and be good, because everybody else is kind of okay. But, like.
563 01:05:55.200 ⇒ 01:05:56.609 Uttam Kumaran: I was like, okay…
564 01:05:56.790 ⇒ 01:06:01.710 Uttam Kumaran: But I also knew how much money was being made, and I was like, this is, like, kind of crazy.
565 01:06:01.860 ⇒ 01:06:13.149 Uttam Kumaran: But I also know that consulting is difficult. You’re almost… you think about it, we’re… we’re like the fractional or the full data team for, like, 15, 20 different companies simultaneously.
566 01:06:13.480 ⇒ 01:06:16.129 Uttam Kumaran: That’s, like, kind of a little bit of a challenge.
567 01:06:16.240 ⇒ 01:06:25.390 Uttam Kumaran: And so the other alternative was to go to, like, a really big company, like a Google or something, and if anybody that knows me knows, like, that would really be a hard place for me to, like.
568 01:06:25.710 ⇒ 01:06:40.019 Uttam Kumaran: I would just, like, make… not make a lot of friends there. Like, you know, I tend to just try to do the things that work and grow, and I’m not… I don’t really like heavy politics environments. At least politics where it’s not, like.
569 01:06:40.420 ⇒ 01:06:59.920 Uttam Kumaran: you know, for, like, it’s… and that’s not politics, though, you know? I think politics were where you have to convince people that solutions are right, or… so I’m not really… I don’t like those types of environments. And so, the third option was just to try this out, but I never… I didn’t start this business being, like, as I was gonna do it as a freelancer. I always was like, we’re gonna start a business.
570 01:07:00.090 ⇒ 01:07:07.669 Uttam Kumaran: But we didn’t raise any money. I’ve been here in Austin for about 4 years, and I quit my job, and then sort of…
571 01:07:07.800 ⇒ 01:07:24.449 Uttam Kumaran: was like, okay, I’ll give myself, like, 3 months to just, like, see if I can get some clients, and then I got our first client, like, in a few months, and then in December of 2023, got our second client, made our first, like, part-time hire in December, and sort of just, like, climbed and climbed.
572 01:07:24.780 ⇒ 01:07:29.680 Uttam Kumaran: You know, now we have about 25 people, sort of scattered everywhere in the world.
573 01:07:29.970 ⇒ 01:07:32.489 Uttam Kumaran: And it’s just…
574 01:07:32.640 ⇒ 01:07:46.629 Uttam Kumaran: it’s kind of the same thing I’ve always been doing, is like, you… you lean on long-term relationships you’ve made, you try to deliver more value, way more value for the client, almost to the point where they’re like, this is cheap, like, these guys are cheap.
575 01:07:46.630 ⇒ 01:08:00.050 Uttam Kumaran: You know? And you always try to do that until the point at which, okay, like, we have a brand, we have a reputation, and the ball starts to roll, and that’s kind of the moment we’re in now.
576 01:08:00.250 ⇒ 01:08:05.150 Uttam Kumaran: I never thought I would be in consulting or professional services. I had very little…
577 01:08:05.420 ⇒ 01:08:19.930 Uttam Kumaran: like, respect for the industry, I feel like, just because I have a lot of friends that are at… I mean, these companies are huge, like, so even if you say, like, I’m at Accenture, you don’t know what team you’re on or whatever, but I just don’t think that, from a tech… from a technologist standpoint.
578 01:08:20.330 ⇒ 01:08:29.260 Uttam Kumaran: those were the companies I ever looked up to, but now that I’m in… I’m in Brainforge, and I see it every day, there are a lot of interesting challenges, like…
579 01:08:29.670 ⇒ 01:08:33.650 Uttam Kumaran: the engineering of our own company is really interesting. We have…
580 01:08:33.920 ⇒ 01:08:50.110 Uttam Kumaran: we have done a lot of things the traditional way. As I mentioned, also, we’ve done things non-traditional. We don’t have, like, a PMO. Like, I skipped… I tried to skip a lot of that. I was a product manager for a while, and so we… so we, like, we did some things our own way, all in the benefit of our clients and our team.
581 01:08:50.720 ⇒ 01:09:04.720 Uttam Kumaran: And then the other thing is, I think of Brainforge ultimately as, like, a platform for everybody that’s in it. Like, we’re a two-sided marketplace. We have a demand, which is coming from clients. We also have the supply, which is, like, our people.
582 01:09:05.069 ⇒ 01:09:11.460 Uttam Kumaran: And Brain Forge, our job is to connect those people, and deliver outcomes for both parties.
583 01:09:11.550 ⇒ 01:09:27.599 Uttam Kumaran: And hopefully we take a little bit in the middle. And I think that’s, like, sort of crude and, like, minimal, but, like, I’ve always had this sort of platform approach to our business. So internally, it’s not like there’s 10 clients and 10 teams and nobody talks to each other. It’s, like, everybody knows each other.
584 01:09:27.680 ⇒ 01:09:38.679 Uttam Kumaran: like, everybody’s able to call on someone who, like, is so far away from their client, or whatever, and still have interactions. So internally, we’re like a product startup company, you know? It’s…
585 01:09:38.729 ⇒ 01:09:57.760 Uttam Kumaran: you wouldn’t… you would… you would think our product is just a service, but internally, it’s not, like, as isolated as I’ve seen many consultancies. And then as soon as we could, we invested in bringing on people, you know, to spend full time here, versus, like, just have a mixed bag of contractors and freelancers.
586 01:09:59.580 ⇒ 01:10:13.560 Uttam Kumaran: In terms of the mission, I don’t know, it’s hard. Like, I just wanted to survive for a long time, still feel like that, because this is kind of crazy that the business, you know, still exists. But I… for me, like, the only thing I think about all the time is just…
587 01:10:13.750 ⇒ 01:10:27.279 Uttam Kumaran: hey, there’s people like Demi and Awash that are spending a year of their career here, they’re extremely talented, and how do I, do right by them and give them a place where they’re, like, pumped to come do work, which, you know, work is work.
588 01:10:27.540 ⇒ 01:10:35.499 Uttam Kumaran: so how can I make it a place where they’re learning, that we’re investing in people? And, again, really, truly, like, is this a place that 3 years ago I would have been, like.
589 01:10:35.930 ⇒ 01:10:37.739 Uttam Kumaran: I’d love to work here. So…
590 01:10:38.050 ⇒ 01:10:51.909 Uttam Kumaran: It’s not… it’s not grown very much from what the original mission was, was to, like, win, to be, like, a really, sophisticated player in the consulting world, to punch up.
591 01:10:52.750 ⇒ 01:10:57.810 Uttam Kumaran: But to also, like, skip all the politics, like, kind of, like, actually have smart people and…
592 01:10:58.000 ⇒ 01:11:03.630 Uttam Kumaran: And deliver outcomes, you know, and just focus on that, like, don’t play any other… Pro-term games.
593 01:11:04.190 ⇒ 01:11:05.830 Hussein Diab: Yeah. Love that.
594 01:11:06.820 ⇒ 01:11:09.910 Uttam Kumaran: What do you think… what do you think about that? Or how does that… how does that, like…
595 01:11:10.030 ⇒ 01:11:11.589 Uttam Kumaran: When you hear that, what do you think?
596 01:11:12.650 ⇒ 01:11:15.769 Hussein Diab: So this is, this is really important, because…
597 01:11:16.200 ⇒ 01:11:20.940 Hussein Diab: this is the whole reason, if… if I were to join you.
598 01:11:21.430 ⇒ 01:11:25.370 Hussein Diab: We have to make sure that our values align.
599 01:11:25.500 ⇒ 01:11:30.430 Hussein Diab: Yeah. You know, we are there going after the same thing.
600 01:11:30.890 ⇒ 01:11:37.459 Hussein Diab: if… If we’re not, if we’re not in alignment, you know, it’s not a win-win anymore.
601 01:11:37.680 ⇒ 01:11:43.929 Hussein Diab: So, that’s why I wanted to ask you this question, and I wanted to make sure. My…
602 01:11:44.320 ⇒ 01:11:53.069 Hussein Diab: My vision aligned with yours. My vision is to help people Get to better places.
603 01:11:53.300 ⇒ 01:11:53.810 Uttam Kumaran: Yeah.
604 01:11:53.810 ⇒ 01:11:57.830 Hussein Diab: Whether it’s internally and on the client side.
605 01:11:57.830 ⇒ 01:12:09.449 Uttam Kumaran: Yes. Both, yeah, totally. Like, this is where I think a lot of people get mistaken, where they’re like, oh… I say clients are number one. Without clients, we don’t have a business. But the team is, like, basically at the same rung.
606 01:12:09.710 ⇒ 01:12:22.210 Uttam Kumaran: there’s… that’s the product. Like, so it’s… it’s very simple for me, you know, and it’s a win-win. I show you try to make everything we play, we try to say, how do we make it a win-win situation? And…
607 01:12:22.320 ⇒ 01:12:30.420 Uttam Kumaran: I think I’ve been… because of my background in… in venture-backed startups, like, I’ve been around a lot of, like, ego, greed.
608 01:12:30.580 ⇒ 01:12:39.659 Uttam Kumaran: that stuff in business, and it sucks. It’s the worst. I don’t like those people. We don’t partner with those people. We don’t hire those people. They don’t make it into the company.
609 01:12:39.790 ⇒ 01:12:52.210 Uttam Kumaran: That’s it. Like… so, which is great, because I’ve been part of a lot of organizations where you’re like, why did we hire this person? Like, you know, how did this happen? And… or, like, their leaders…
610 01:12:52.760 ⇒ 01:13:00.950 Uttam Kumaran: sort of on a yacht somewhere all the time. I don’t like… I’m not… that’s not… that’s not it. Like, that’s… I can’t build a business like that, so…
611 01:13:03.170 ⇒ 01:13:03.880 Hussein Diab: Awesome.
612 01:13:04.050 ⇒ 01:13:04.720 Hussein Diab: Well…
613 01:13:05.000 ⇒ 01:13:14.589 Hussein Diab: Yeah, that’s all I have for now, that’s all I have for the questions. I appreciate the time, I appreciate staying extra few minutes to help me answer these questions.
614 01:13:14.590 ⇒ 01:13:24.069 Uttam Kumaran: Definitely. Thank you, Hussein. Sorry for the delay earlier. So, yeah, and I think we’ll discuss internally, and we’ll… we’ll discuss with Kayla, and, I’m sure you’ll be hearing back soon.
615 01:13:24.450 ⇒ 01:13:26.670 Hussein Diab: Cool. Awesome. Nice to meet you, Autumn.
616 01:13:26.670 ⇒ 01:13:28.130 Uttam Kumaran: Alright, thank you. Thank you. Bye-bye.
617 01:13:28.130 ⇒ 01:13:29.510 Demilade Agboola: Thank you. Bye.