Meeting Title: Brainforge Analytics Engineer Interview Date: 2026-03-16 Meeting participants: Awaish Kumar, Mateo
WEBVTT
1 00:01:05.820 ⇒ 00:01:06.690 Awaish Kumar: Here we are.
2 00:01:15.290 ⇒ 00:01:16.339 Awaish Kumar: That sounds right.
3 00:01:23.320 ⇒ 00:01:24.090 Mateo: Hello?
4 00:01:24.350 ⇒ 00:01:26.110 Mateo: Good afternoon, can you hear me?
5 00:01:27.120 ⇒ 00:01:28.530 Awaish Kumar: Yes, I can.
6 00:01:29.340 ⇒ 00:01:30.989 Awaish Kumar: Hi, how you doing?
7 00:01:31.630 ⇒ 00:01:33.390 Mateo: I’m doing great, how about you?
8 00:01:33.890 ⇒ 00:01:36.510 Awaish Kumar: I’m good. Where are you located?
9 00:01:37.750 ⇒ 00:01:41.120 Mateo: I am based in Colombia. What about you?
10 00:01:41.930 ⇒ 00:01:42.620 Awaish Kumar: Sorry?
11 00:01:43.270 ⇒ 00:01:44.390 Mateo: What about you?
12 00:01:45.760 ⇒ 00:01:47.870 Awaish Kumar: Yeah, I didn’t hurl, like, where are you?
13 00:01:48.360 ⇒ 00:01:49.120 Awaish Kumar: Rome?
14 00:01:49.800 ⇒ 00:01:53.010 Mateo: Right, yeah, from Colombia, Latin America.
15 00:01:53.010 ⇒ 00:01:55.499 Awaish Kumar: Lovely, okay, I’m right now enjoy…
16 00:01:57.370 ⇒ 00:01:58.800 Mateo: In Word? I’m sorry?
17 00:01:58.800 ⇒ 00:01:59.630 Awaish Kumar: Joy.
18 00:02:00.490 ⇒ 00:02:02.410 Mateo: Okay, nice, nice, nice.
19 00:02:03.560 ⇒ 00:02:13.500 Awaish Kumar: Okay, so basically the agenda for today’s interview is just to get to know each other, and understand more about your background and experience.
20 00:02:14.130 ⇒ 00:02:22.929 Awaish Kumar: And how can… Basically, you’ve… you can fit in the role we have. Head brain forge,
21 00:02:23.110 ⇒ 00:02:29.479 Awaish Kumar: So basically, yeah, before I dive into, like, the specific Things? Like, do…
22 00:02:29.720 ⇒ 00:02:33.020 Awaish Kumar: Can you brief me, like, if you know anything about Brainforge or anything?
23 00:02:33.920 ⇒ 00:02:41.090 Mateo: Well, just… some things I have, like, quickly spot-checked your…
24 00:02:41.320 ⇒ 00:02:46.060 Mateo: The website, did some research further. I know that you…
25 00:02:46.310 ⇒ 00:02:54.079 Mateo: Guys basically consult our data, and, like, develop projects, really.
26 00:02:54.210 ⇒ 00:02:56.810 Mateo: data-driven projects, AI, cloud.
27 00:02:57.270 ⇒ 00:03:00.340 Mateo: But, yeah, that’s… that’s… that’s about it.
28 00:03:01.400 ⇒ 00:03:06.490 Awaish Kumar: Okay, okay. And, okay, so…
29 00:03:07.310 ⇒ 00:03:13.859 Awaish Kumar: where, like, I see your, like, past notes I have from the… from your loom, and also…
30 00:03:14.190 ⇒ 00:03:15.829 Awaish Kumar: From your profile.
31 00:03:15.970 ⇒ 00:03:19.410 Awaish Kumar: It says… it mentions, like, both the things that you…
32 00:03:19.720 ⇒ 00:03:31.889 Awaish Kumar: you have worked as an analytics engineer, and also you’re kind of doing data engineering right now? So where do you see yourself in the next role? Working as an analytics engineer, or a data engineer?
33 00:03:32.880 ⇒ 00:03:37.679 Mateo: Yeah, I see… I see that this… this analytics engineering role has really…
34 00:03:37.920 ⇒ 00:03:46.010 Mateo: blossom bloomed over the past years, because right now, with the… with AI, companies are…
35 00:03:46.250 ⇒ 00:03:52.760 Mateo: Maybe more reclutant in spending a lot of money in their data teams and organizations, so…
36 00:03:52.930 ⇒ 00:04:10.969 Mateo: most of the times, it’s just a couple guys, or just one person handling the analytics process, especially for small companies, from beginning to end, so the analytics engineer role that takes care of the whole ETL transformation, polishing, and also
37 00:04:11.300 ⇒ 00:04:20.109 Mateo: visualization and insights, it’s really appreciated. So yeah, I think that it’s a great role, and I want to continue down that path.
38 00:04:20.529 ⇒ 00:04:21.360 Mateo: It’s great.
39 00:04:22.440 ⇒ 00:04:23.890 Awaish Kumar: Okay,
40 00:04:24.140 ⇒ 00:04:43.889 Awaish Kumar: Yeah, like, I understand what you just said, like, now that one role is doing a lot of things, especially with AI, and also if you’re in a small company, but also, like, in our company, we have different work streams, right? We have data engineering, analytics engineering, data analysts.
41 00:04:44.080 ⇒ 00:04:48.049 Awaish Kumar: So, that’s why I just want to clarify, like, where you wanna…
42 00:04:48.320 ⇒ 00:04:52.780 Awaish Kumar: But where you want to be when you are in the rainforest, basically.
43 00:04:52.780 ⇒ 00:04:57.179 Mateo: Right, right, right, right. Yeah, the… the whole analytics engineer…
44 00:04:57.320 ⇒ 00:04:59.979 Mateo: position, it works really good, yeah.
45 00:04:59.980 ⇒ 00:05:00.600 Awaish Kumar: Okay.
46 00:05:00.770 ⇒ 00:05:06.150 Awaish Kumar: And, yeah, so how,
47 00:05:06.490 ⇒ 00:05:10.949 Awaish Kumar: Do you rate yourself, in terms of a Python SQL?
48 00:05:11.690 ⇒ 00:05:12.610 Awaish Kumar: In terms of…
49 00:05:12.610 ⇒ 00:05:13.209 Mateo: I don’t…
50 00:05:13.210 ⇒ 00:05:14.569 Awaish Kumar: How would you rate yourself?
51 00:05:15.810 ⇒ 00:05:18.119 Mateo: Right, well, I…
52 00:05:18.480 ⇒ 00:05:35.180 Mateo: like before, in the past years, I, like, did some heavily coding with SQL and Python, but lately, I’ve been using AI a lot. It makes my day easier, so I don’t, like, spend a lot of time on
53 00:05:35.430 ⇒ 00:05:41.170 Mateo: On coding and building that stuff, but overseeing, basically, how…
54 00:05:41.380 ⇒ 00:05:45.820 Mateo: How things are going, performance, costs.
55 00:05:46.780 ⇒ 00:05:52.499 Mateo: insights. So, so yeah, but to give you a number, I believe… 8.
56 00:05:53.370 ⇒ 00:05:54.000 Awaish Kumar: Right.
57 00:05:55.010 ⇒ 00:06:09.350 Awaish Kumar: Yeah, I understand that, and with… with the, like, with the VIP coding right now, everybody uses AI to write the code, that’s okay, but I just… I want to know that, like, if you know the depth of these languages, if…
58 00:06:09.650 ⇒ 00:06:26.510 Awaish Kumar: AI spits out some code which you… would you be able to identify, like, if… if this is not optimal, or there are any issues, and things like that. So, like, like, for example, I have one of the questions, if you have a…
59 00:06:26.660 ⇒ 00:06:42.389 Awaish Kumar: for example, I have a table which has, like, billions of rows, and then I want to curate, and it takes, like, 5 to 10 minutes to execute my query. That’s very slow, right? I know that it can be optimized under maybe 30 seconds, so…
60 00:06:42.970 ⇒ 00:06:47.269 Awaish Kumar: What steps you would take, to optimize that?
61 00:06:48.430 ⇒ 00:07:00.789 Mateo: Right, well, first of all, it depends. Right now, for example, if you’re using Snowflake for cloud and for analytics recording, it has
62 00:07:00.790 ⇒ 00:07:12.310 Mateo: their own, like, clustering and a micro-partition optimization, which is native. So you just tweak maybe the primary key that Snowflake needs to check, and
63 00:07:12.540 ⇒ 00:07:31.979 Mateo: the query will be optimized. Also, you can scale horizontal and vertical on use more resources if needed. It’s really complicated. But there are some techniques, especially… I was working with SQL Server, it was the whole indexing, which is basically what Snowflake does.
64 00:07:32.050 ⇒ 00:07:41.469 Mateo: Natively, so of course, you can rely on those things. Also, if there’s joins, you have to check the…
65 00:07:41.770 ⇒ 00:07:48.630 Mateo: The filtering and the… basically the order that the… the… the query is… is being…
66 00:07:49.670 ⇒ 00:07:59.289 Mateo: So in Perform, so you know on which steps do you narrow the amount of bytes that it needs to scan, so you can optimize. But yeah, it depends.
67 00:08:01.750 ⇒ 00:08:08.459 Awaish Kumar: Yeah, so Okay, so is, like, have you used tools like dbt in your work?
68 00:08:09.300 ⇒ 00:08:16.810 Mateo: Yes, yes. Lately, I’ve been using DVT a lot for the whole transformation Oh.
69 00:08:16.810 ⇒ 00:08:23.470 Awaish Kumar: So, for example, if it was a dbt model, so what changes you would make?
70 00:08:25.110 ⇒ 00:08:28.760 Mateo: Right, well, you will need to check the joins.
71 00:08:28.870 ⇒ 00:08:34.890 Mateo: For sure, referencing, so you…
72 00:08:35.520 ⇒ 00:08:44.280 Mateo: you almost want… almost always want to reference models that are already there, Liberty and not… like, external…
73 00:08:44.550 ⇒ 00:09:00.099 Mateo: tables, because if you’re referencing external tables, then you rely on the computing from the cloud services that you’re using, so you will need to check that, but if it’s everything that’s on DVT, you can rely on… on joins,
74 00:09:00.610 ⇒ 00:09:03.359 Mateo: Well, maybe indexing will be also good.
75 00:09:03.490 ⇒ 00:09:08.180 Mateo: And yeah, basically, what kind of…
76 00:09:08.600 ⇒ 00:09:14.120 Mateo: How are you modeling data? If you’re using just a single
77 00:09:14.450 ⇒ 00:09:21.469 Mateo: Fuck thing to avoid joints, or if you’re using a snowflake schema to To avoid computing resources.
78 00:09:21.670 ⇒ 00:09:24.820 Mateo: But yeah, I think most of the times I just use the…
79 00:09:24.940 ⇒ 00:09:29.680 Mateo: One whole big table, and deal with that much data, but… but yeah.
80 00:09:29.800 ⇒ 00:09:31.629 Mateo: Here are some techniques, too.
81 00:09:31.630 ⇒ 00:09:34.640 Awaish Kumar: You mentioned about Snowflake schema, so, like,
82 00:09:34.880 ⇒ 00:09:38.580 Awaish Kumar: What is the difference between snowflake and a star schema?
83 00:09:39.770 ⇒ 00:09:45.509 Mateo: Right, it’s, basically the denormalization, if you have the…
84 00:09:46.310 ⇒ 00:09:49.920 Mateo: The star schema, you have a fact table, big one, and…
85 00:09:50.070 ⇒ 00:09:56.579 Mateo: And your whole dimension tables, and of course, there will be, like, data that will repeat.
86 00:09:57.260 ⇒ 00:09:59.880 Mateo: Every table to be able to join to the…
87 00:10:00.250 ⇒ 00:10:11.150 Mateo: To the main one, so it may be… you may be wary of a computer store, which may be not that significant, but something to think about, because…
88 00:10:11.390 ⇒ 00:10:13.139 Mateo: It will store a lot of…
89 00:10:13.350 ⇒ 00:10:20.460 Mateo: redundant data, and there’s a snowflake where you basically query based on, joins and…
90 00:10:21.250 ⇒ 00:10:24.479 Mateo: And basically, yeah, they denormalize things.
91 00:10:29.410 ⇒ 00:10:30.490 Awaish Kumar: Okay.
92 00:10:30.670 ⇒ 00:10:34.790 Awaish Kumar: And, then how… yeah, so…
93 00:10:37.280 ⇒ 00:10:42.209 Awaish Kumar: What is, like, normalization? What is denormalization?
94 00:10:44.010 ⇒ 00:10:51.900 Mateo: I may be confusing normalization with denormalization, but, normalized tables are basically the…
95 00:10:52.140 ⇒ 00:10:55.029 Mateo: The star one, in which you repeat.
96 00:10:55.170 ⇒ 00:10:58.610 Mateo: Data, which you have the big one, and then…
97 00:10:59.070 ⇒ 00:11:07.590 Mateo: your associated tables, and denormalization is when you have, for example, this snowflake schema where you don’t have
98 00:11:08.040 ⇒ 00:11:11.370 Mateo: Your whole reductant data, so…
99 00:11:11.720 ⇒ 00:11:19.800 Mateo: There’s joints and more complicated relationship between the table, and you break it down into more… Small chunks, yep.
100 00:11:21.790 ⇒ 00:11:27.479 Awaish Kumar: Okay, I just have a few more questions regarding,
101 00:11:27.790 ⇒ 00:11:35.179 Awaish Kumar: How would you communicate, with the non-technical stakeholders if they disagree with your…
102 00:11:36.070 ⇒ 00:11:45.280 Awaish Kumar: findings or whatever you showed them, for example? And if they don’t… if they disagree, so how would you make, your… your…
103 00:11:45.560 ⇒ 00:11:48.979 Awaish Kumar: analysis, or bakery, or what you’ve… what you’re saying.
104 00:11:50.670 ⇒ 00:11:54.069 Mateo: Right, I think backing up everything that you…
105 00:11:54.430 ⇒ 00:12:06.400 Mateo: show or say with data, it’s, of course, really, really important, so if you’re saying that X or Y variable is dropping down because of this, you need to show.
106 00:12:06.520 ⇒ 00:12:09.260 Mateo: Basically, the… the relationship and…
107 00:12:09.770 ⇒ 00:12:25.780 Mateo: and explain yourself with… visually, it will be better, because usually people are really visual, and they tend to consume information and data based on graphs and visuals and whatever, but I think that the most
108 00:12:26.380 ⇒ 00:12:34.810 Mateo: Shocking part is when they… Oh… like… the…
109 00:12:34.950 ⇒ 00:12:38.910 Mateo: They don’t know in depth, like you mentioned, they’re not technical.
110 00:12:38.980 ⇒ 00:12:51.130 Mateo: person, so if they want to add, like, hey, just, we need to add a column, or we need to, like, filter out this, that may sound really simple for them, like…
111 00:12:51.160 ⇒ 00:13:00.080 Mateo: Just to let this, but there’s a lot of complicated background work when you need to add something or delete something, so you need to…
112 00:13:00.090 ⇒ 00:13:14.389 Mateo: Basically explain step-by-step what is needed to achieve the solution, and give them a really clear vision of what you’re doing, what they can expect, and…
113 00:13:14.570 ⇒ 00:13:17.610 Mateo: What will be the work, Roland, what you’re asking?
114 00:13:21.530 ⇒ 00:13:28.799 Awaish Kumar: Okay, and, so how, like, can you give me an example of,
115 00:13:29.220 ⇒ 00:13:32.639 Awaish Kumar: Of a tool that you learned during… during the job?
116 00:13:32.870 ⇒ 00:13:38.390 Awaish Kumar: And basically use that to… To… To finish, like, complete something.
117 00:13:40.620 ⇒ 00:13:49.029 Mateo: Yeah, a lot of… a lot of things I have learned while working. One of them was… was DVT.
118 00:13:49.230 ⇒ 00:13:50.730 Mateo: There are some…
119 00:13:51.390 ⇒ 00:14:01.299 Mateo: like, specific industry versions of variations of certain tools. For example, Fivetran is the standard for extraction, but
120 00:14:01.960 ⇒ 00:14:05.750 Mateo: Marketing-based companies use…
121 00:14:05.950 ⇒ 00:14:15.149 Mateo: at Verity, for example, for data extraction, which is focused for marketing tools, so you have to learn that on the way. You have your
122 00:14:15.350 ⇒ 00:14:23.630 Mateo: like everything, you have your… your solid bases. For example, when you switch from cloud to cloud, so I was used to
123 00:14:24.010 ⇒ 00:14:40.620 Mateo: using Snowflake, but some companies use GCP, so the UI and the options change, but the idea and the logic behind it is pretty much the same, so I think that if you have, like, solid
124 00:14:41.140 ⇒ 00:14:45.030 Mateo: Fundamentals and concepts, and you understand what’s
125 00:14:45.300 ⇒ 00:14:51.330 Mateo: The idea behind, you can pretty much master every variation of the tool that you’re using.
126 00:14:53.290 ⇒ 00:14:53.950 Awaish Kumar: Okay.
127 00:14:54.530 ⇒ 00:15:00.309 Awaish Kumar: Yeah, I think I’m good with my questions. Like, do you have any… Questions for me?
128 00:15:01.350 ⇒ 00:15:07.870 Mateo: Well, yeah, I don’t know if you are aware of the position, or…
129 00:15:08.340 ⇒ 00:15:12.790 Mateo: or the client, if I can ask you some questions about it, or…
130 00:15:12.970 ⇒ 00:15:13.770 Awaish Kumar: Again, you can.
131 00:15:14.500 ⇒ 00:15:21.399 Mateo: Okay, great. So I will… Be interested in knowing a bit more about the day-to-day.
132 00:15:21.580 ⇒ 00:15:24.290 Mateo: Activities of, of, of the role, and…
133 00:15:24.500 ⇒ 00:15:28.049 Awaish Kumar: Yeah, day-to-day activities are, like, just like,
134 00:15:28.180 ⇒ 00:15:35.900 Awaish Kumar: Standout meeting with the… with your team, planning out the day, what you will be delivering by the end of day.
135 00:15:36.050 ⇒ 00:15:41.260 Awaish Kumar: And, maybe on Mondays, you meet and you discuss about the whole week.
136 00:15:41.680 ⇒ 00:15:45.640 Awaish Kumar: And then, on the… basically…
137 00:15:46.050 ⇒ 00:15:50.159 Awaish Kumar: people communicate in Slack, that is, like, kind of us in communication.
138 00:15:50.280 ⇒ 00:15:54.140 Awaish Kumar: Between team members, you can write document, we have… we use linear.
139 00:15:54.270 ⇒ 00:15:57.489 Awaish Kumar: As a project management tool, you can use that.
140 00:15:57.950 ⇒ 00:16:00.800 Awaish Kumar: Communicate in Slack with your team members.
141 00:16:01.050 ⇒ 00:16:08.179 Awaish Kumar: Drive the… the delivery of the work you have been assigned. And basically, for that, if you want to…
142 00:16:08.320 ⇒ 00:16:12.230 Awaish Kumar: If you have to meet with anyone, you are free to just Slack them.
143 00:16:12.540 ⇒ 00:16:18.490 Awaish Kumar: And, yeah, we focus on writing documentation so that… or creating
144 00:16:18.700 ⇒ 00:16:27.319 Awaish Kumar: short looms, basically, if you want to ask something, you can record a loom and send it in the channel to ask. So these are basically…
145 00:16:27.640 ⇒ 00:16:34.080 Awaish Kumar: The ways of communication, and the stand-up is something that we start off our day with.
146 00:16:34.270 ⇒ 00:16:37.959 Awaish Kumar: And apart from that, like, you might be assigned
147 00:16:38.760 ⇒ 00:16:42.160 Awaish Kumar: On, more than one client, so we have,
148 00:16:42.410 ⇒ 00:16:53.950 Awaish Kumar: clients, multiple clients here, and not everybody needs, like, 40 hours of an analytics engineer. So we might have to split your time between maybe 2 to 3 clients, where…
149 00:16:54.260 ⇒ 00:17:07.339 Awaish Kumar: You will be, like, managing your time, on a daily basis, like, working between one or two clients, so maybe 3-4 hours on a one client, and then 3-4 hours on a second client, and something like that.
150 00:17:07.770 ⇒ 00:17:11.889 Awaish Kumar: And, yeah, that’s basically how it is.
151 00:17:12.690 ⇒ 00:17:24.219 Mateo: Okay, sounds great, so it’s really dynamic, so, do you guys have, like, a good community, and you guys help each other? Like, you have some Slack-dedicated channels, if anyone has.
152 00:17:24.220 ⇒ 00:17:33.779 Awaish Kumar: Yeah, a lot of Slack channels, help channels, also, like, team members. Everybody’s really open. If you ask any questions, like, people will reply.
153 00:17:35.400 ⇒ 00:17:51.580 Awaish Kumar: on your, for example, whatever question is, and then if you record a loom, that will be much more, like, easier for anybody to help. Apart from that, also, like, you can huddle, like, call anyone, and
154 00:17:52.230 ⇒ 00:17:59.000 Awaish Kumar: like, get around whatever you’re working. So it’s really kind of a helpful Team,
155 00:17:59.630 ⇒ 00:18:04.609 Awaish Kumar: Yeah, so yeah, so that’s how, in the… in the team, like, in the… like, we are not…
156 00:18:04.860 ⇒ 00:18:07.520 Awaish Kumar: I’m planning for this new startup, and we are, like.
157 00:18:07.690 ⇒ 00:18:13.119 Awaish Kumar: A close team of, like, in terms of engineering, we are, like, maybe 10 to 15 people.
158 00:18:13.260 ⇒ 00:18:19.740 Awaish Kumar: Between, AI and data… data engineering, analysts, maybe 15 to 20, maybe.
159 00:18:20.010 ⇒ 00:18:23.039 Awaish Kumar: And, so it’s easy to basically
160 00:18:23.370 ⇒ 00:18:31.229 Awaish Kumar: remember everyone, and also there are specific… specific, select channels, like for the AI team, data team.
161 00:18:31.500 ⇒ 00:18:34.489 Awaish Kumar: So you can go in whatever help you need, for example.
162 00:18:34.670 ⇒ 00:18:38.810 Awaish Kumar: You are working for a data client, and your client might need some…
163 00:18:39.350 ⇒ 00:18:51.959 Awaish Kumar: AI services. So, obviously, you don’t know about those things, you can ask an AI channel, okay, my cloud is asking for that, how can I do that? Or… or if they can be able to help you, so this is how we…
164 00:18:52.390 ⇒ 00:18:53.650 Awaish Kumar: Like, help each other.
165 00:18:54.420 ⇒ 00:18:57.889 Mateo: Okay, sounds great, sounds great. Thank you very much for the…
166 00:18:58.150 ⇒ 00:19:00.160 Mateo: For the answers, and for the time.
167 00:19:00.940 ⇒ 00:19:03.990 Awaish Kumar: Okay, yeah, no worries. Anything else?
168 00:19:04.960 ⇒ 00:19:12.610 Mateo: No, I think we’re good. We’re good. Thank you very much. I don’t know if you know anything about next steps, or…
169 00:19:12.610 ⇒ 00:19:13.130 Awaish Kumar: I’m checking.
170 00:19:13.130 ⇒ 00:19:13.660 Mateo: Right.
171 00:19:13.850 ⇒ 00:19:27.370 Awaish Kumar: I submit my feedback, maybe the team will reach out to you, regarding the next steps, but normally our next steps are, like, a second interview with one of my colleagues, and
172 00:19:27.650 ⇒ 00:19:32.439 Awaish Kumar: Then, maybe take-home assignment, and a panel interview.
173 00:19:32.900 ⇒ 00:19:36.180 Awaish Kumar: And after that, there will be an offer.
174 00:19:36.350 ⇒ 00:19:37.179 Awaish Kumar: That’s all.
175 00:19:38.100 ⇒ 00:19:40.570 Mateo: That’s good, yeah, really good. Thank you for that.
176 00:19:41.180 ⇒ 00:19:42.510 Awaish Kumar: Okay, thank you.
177 00:19:43.500 ⇒ 00:19:44.689 Mateo: Thank you, you have an easy…
178 00:19:44.940 ⇒ 00:19:45.640 Awaish Kumar: Yep.