Meeting Title: Brainforge Interview w- Demilade Date: 2026-04-20 Meeting participants: Nikhil G, Demilade Agboola
WEBVTT
1 00:03:05.660 ⇒ 00:03:07.030 Demilade Agboola: Hi.
2 00:03:07.030 ⇒ 00:03:07.920 Nikhil G: Hey.
3 00:03:08.170 ⇒ 00:03:09.740 Demilade Agboola: Hi.
4 00:03:09.950 ⇒ 00:03:12.000 Demilade Agboola: I’m doing very well, how are you?
5 00:03:12.000 ⇒ 00:03:15.059 Nikhil G: Good, good, yeah, no, all good.
6 00:03:15.060 ⇒ 00:03:30.680 Demilade Agboola: That’s good to hear. So my name is Dami Lade. I work with the Brainforge team as a Senior Analytics Engineer, and I will be here to just talk to you, get your idea on how you
7 00:03:30.970 ⇒ 00:03:34.330 Demilade Agboola: Walk through systems, and how you design systems.
8 00:03:34.480 ⇒ 00:03:48.320 Demilade Agboola: And that would just basically be the concept of this interview. So there won’t be any, like, you know, live coding or anything of that sort. I just want to understand, you, your approach, and how you solve, problems.
9 00:03:49.670 ⇒ 00:03:51.300 Nikhil G: Sounds good, sounds good, yeah, yeah.
10 00:03:51.300 ⇒ 00:03:55.559 Demilade Agboola: Okay. Alright, just to start off, how do I pronounce your name, please?
11 00:03:56.140 ⇒ 00:03:57.570 Nikhil G: It’s, Nikhil? Yeah.
12 00:03:57.570 ⇒ 00:03:59.160 Demilade Agboola: We do. Okay.
13 00:03:59.160 ⇒ 00:04:00.270 Nikhil G: How about you?
14 00:04:00.270 ⇒ 00:04:01.590 Demilade Agboola: Dimiladi.
15 00:04:02.010 ⇒ 00:04:03.990 Nikhil G: Damien Lady, okay. Alright.
16 00:04:04.210 ⇒ 00:04:08.469 Demilade Agboola: Nice to meet you, too. Alright, so can we…
17 00:04:08.650 ⇒ 00:04:16.590 Demilade Agboola: get started. Can you please tell me about yourself, and your technical, stack?
18 00:04:17.510 ⇒ 00:04:19.440 Nikhil G: Yeah, sure, definitely. So…
19 00:04:19.560 ⇒ 00:04:31.539 Nikhil G: Yeah, like, I’m, working in data industry for more than 11 years now, you know, like, I have started my career with the IBM, and then moved to different,
20 00:04:31.600 ⇒ 00:04:49.980 Nikhil G: companies work on the different tech stack, being from the Hadoop on-premises systems in the last 5 to 6 years, more focused towards the cloud solutions, building data warehouses in the cloud, using Snowflake, Databricks, you know, Redshift, and all that sort of stuff, so…
21 00:04:49.980 ⇒ 00:05:09.539 Nikhil G: heavily used the Python, all types of SQLs, you know, all the databases, like the MySQL, Oracle, and all that sort of legacy thing, and as well as the… recently, in the last, I think, 2 to 4 years, like, heavily working on the Airflow and dbt, you know, to build the transformation layers.
22 00:05:09.540 ⇒ 00:05:13.990 Nikhil G: Whenever the data is ingested into,
23 00:05:14.210 ⇒ 00:05:25.420 Nikhil G: databases, let’s, talk about the Snowflake, and, you know, like, it’s, it’s, the, the solutions are built mostly around, like, the test-driven development.
24 00:05:25.420 ⇒ 00:05:44.810 Nikhil G: And test-driven, the orchestration as well, you know? So, like, it’s not just, like, setting up the pipeline, but we should be doing all sort of checks on it, you know? So, the data monitoring, data quality, observability, and everything has to be in a single place using the alerting, proactive monitoring, and all that sort of stuff.
25 00:05:44.900 ⇒ 00:05:58.400 Nikhil G: Even said that, I’m also certified in Snowflake, AWS, Airflow, I’m preparing for dbt now, you know, so… Okay. Just like I’m not getting time to…
26 00:05:58.470 ⇒ 00:06:09.039 Nikhil G: more, to prepare and then give exam, but I think, like, hands-on I have heavily work on the DBD, so it will be just, like, complimentary to have that.
27 00:06:09.120 ⇒ 00:06:14.450 Demilade Agboola: Oh, definitely. I actually have the dBT certification, so it’s, it’s, pretty…
28 00:06:14.630 ⇒ 00:06:16.579 Nikhil G: Oh, is it? Okay, thanks, yeah.
29 00:06:16.580 ⇒ 00:06:26.709 Demilade Agboola: And it’s also something we’re trying to also do internally within the team, so that people can get testified across multiple things as well, so that, you know…
30 00:06:26.880 ⇒ 00:06:27.820 Nikhil G: Okay.
31 00:06:28.190 ⇒ 00:06:34.110 Demilade Agboola: we have expertise in-house. Okay, that sounds good. That’s… that’s really nice to hear.
32 00:06:34.330 ⇒ 00:06:38.599 Demilade Agboola: So I think that my next question will be,
33 00:06:38.830 ⇒ 00:06:57.380 Demilade Agboola: how will you… and this is, like, a scenario question, I just want to hear how you think of these different scenarios, your questions, your assumptions, and how you just want to, like, go out building this. So, if we have a client that wants a daily revenue reporting marked.
34 00:06:57.810 ⇒ 00:07:03.589 Demilade Agboola: And they have 3 different sources. They have Stripe, they have Salesforce, and they have Google Ads.
35 00:07:04.930 ⇒ 00:07:07.150 Demilade Agboola: How would you design the solution?
36 00:07:08.230 ⇒ 00:07:10.440 Nikhil G: Sorry, what was the stripe?
37 00:07:10.560 ⇒ 00:07:11.830 Demilade Agboola: Salesforce.
38 00:07:12.230 ⇒ 00:07:13.750 Demilade Agboola: And Google Ads.
39 00:07:13.750 ⇒ 00:07:14.450 Nikhil G: Okay.
40 00:07:15.180 ⇒ 00:07:21.529 Demilade Agboola: how would you design the solution? What are your assumptions? What questions would you want to ask?
41 00:07:21.700 ⇒ 00:07:25.920 Demilade Agboola: And what infrastructure would you choose, and why?
42 00:07:26.820 ⇒ 00:07:41.669 Nikhil G: Yeah, yeah, sounds good, you know, like, definitely, I, I will go for any scalable, more robust, like, the database for, for, like, building this, reporting on, on ingestion. Like, we…
43 00:07:41.670 ⇒ 00:07:53.059 Nikhil G: have to consider the scalability as well, like, people start with the smaller version, but, like, they forget about, like, the what will be the bigger version of it, you know, like, as the company grows and all that sort of stuff, so, you know, so…
44 00:07:53.080 ⇒ 00:08:08.559 Nikhil G: Let’s consider that as well. I will… let’s… let’s… let’s take… we will use Snowflake, you know, sake of simplicity. There are three parts here, like the ingestion, transformations, and then the actual reporting, you know? So, we will keep all three systems differently, decoupled.
45 00:08:08.560 ⇒ 00:08:23.729 Nikhil G: So, ingestion, like, there are, like, plenty of ways now, you know, the open source technologies, like, you can use any sort of tool, you know, like, depending upon client’s budget, how soon they want to get this thing ready, you know? So, they have budget, and they don’t want to,
46 00:08:23.730 ⇒ 00:08:28.419 Nikhil G: they don’t want to, like, reinvent the wheel, you know? Like, we can definitely go with the…
47 00:08:28.420 ⇒ 00:08:39.010 Nikhil G: Fitran, or similar, like, the ingestion, where, like, you can directly connect to your Stripe, Salesforce, and Google Ads, or any funnel, or all sort of stuff.
48 00:08:39.010 ⇒ 00:08:53.020 Nikhil G: And you will directly get that data into Snowflake in Landing Zone. So, we will create 3 parts here, like the medallion architecture, the bronze layer, silver layer, gold layer, you can call it differently, raw base, DW.
49 00:08:53.540 ⇒ 00:09:04.000 Nikhil G: So, FITRAN I will use, if we don’t want to spend on the paid tools, I can go for the data load tool, DLT, which is similar to dbt.
50 00:09:04.080 ⇒ 00:09:23.849 Nikhil G: It’s open source. Using Python, you can simply connect to Stripe, Google AdSense, and other things. We have to just provide the API key, and then all the error handling, API, like the pagination and all that stuff is managed by DLT, you know? Like, it needs some,
51 00:09:23.970 ⇒ 00:09:36.530 Nikhil G: efforts to build that one, but it’s really cool I have implemented that as well, you know? So that is the ingestion part. We will get the data as it is in the Snowflake. We won’t do any transformation during the loading.
52 00:09:36.530 ⇒ 00:09:57.889 Nikhil G: Because we won’t draw, you know, like, so that, like, we can play around that one, we can audit it, what was there, and all that sort of stuff. And from then onwards, like, the actual work starts on the transformation layer, you know? Like, for that, definitely dbt is a frontrunner, you know? Like, there are a few other transformation tools as well, but it makes sense to stick with the dbt, like…
53 00:09:57.930 ⇒ 00:10:17.959 Nikhil G: then you will have the base layer on top of this raw layer, where you will be just, like, fixing the data type corrections, column name correction, table name corrections, you will bring it to the standard formats, you know? So, the customer ID is, like, let’s say it’s a string across all the data sources.
54 00:10:17.960 ⇒ 00:10:25.470 Nikhil G: In one table, it could be, number, worker, but, like, we will bring it to the same format, you know? Like, similarly, we will do…
55 00:10:25.470 ⇒ 00:10:28.239 Nikhil G: all that stuff. And then, like,
56 00:10:28.810 ⇒ 00:10:36.460 Nikhil G: Actually, we will build the… the fact dimensional tables, or, like, we will think about, like, what data modeling we want to do.
57 00:10:36.460 ⇒ 00:10:54.060 Nikhil G: Using dbt, again, like, we will use, like, strategies, materialization, whether we want to use views, tables, incremental models, and all that sort of stuff. We will add a lot of dbt tests, you know, to see, proactively monitor if there are any duplicates, missing data.
58 00:10:54.060 ⇒ 00:11:12.389 Nikhil G: data freshness checks, data anomaly, volume checks, and all that sort of stuff. And then, finally, we will create the data mark. In that one, like, we will have, like, the final report ready, so that, we can directly feed that into your Tableau, or Power BI, or whatever business tool you use.
59 00:11:12.480 ⇒ 00:11:29.999 Nikhil G: And then all this thing, we can orchestrate through Airflow or any other orchestrator that has, like, the event-driven, orchestration, you know? So we won’t set the times for each job, rather, as soon as the data is arrived, we will
60 00:11:30.840 ⇒ 00:11:55.820 Nikhil G: transform it, we will build that one, and then publish it to the tableau, you know, rather than one job runs at 1 o’clock, 2 o’clock, 3 o’clock. If the 1 o’clock job fails, then 2 o’clock, 3 o’clock will still run, it will still produce incorrect data, and all that sort of stuff, you know? So, overall, like, we will build this kind of system, which is more data-aware, and then, towards the solving the actual business problem, rather than just, like.
61 00:11:55.820 ⇒ 00:11:56.490 Nikhil G: data.
62 00:11:56.490 ⇒ 00:12:02.950 Demilade Agboola: I’m sorry to interrupt, there’s someone ringing on my doorbell. Give me one second, I apologize for that.
63 00:12:03.340 ⇒ 00:12:04.030 Nikhil G: Notice.
64 00:13:05.260 ⇒ 00:13:07.659 Demilade Agboola: Alright, turn back.
65 00:13:07.660 ⇒ 00:13:08.590 Nikhil G: I don’t know, chef.
66 00:13:09.940 ⇒ 00:13:13.770 Nikhil G: Totally understand remote working, yeah.
67 00:13:13.770 ⇒ 00:13:19.220 Demilade Agboola: Yeah, you get calls, or you get, like, people ringing at your bell at the weirdest times.
68 00:13:19.220 ⇒ 00:13:21.460 Nikhil G: By the way, where are you dialing from today?
69 00:13:21.790 ⇒ 00:13:26.150 Demilade Agboola: So I’m in Malta, I live in Malta. What about you?
70 00:13:26.150 ⇒ 00:13:28.980 Nikhil G: I’m in Ireland.
71 00:13:29.640 ⇒ 00:13:31.860 Demilade Agboola: Yeah. Thanks, Dublin?
72 00:13:31.860 ⇒ 00:13:37.010 Nikhil G: Dublin, yeah, yeah, must be a good sunny weather there right now.
73 00:13:37.010 ⇒ 00:13:39.899 Demilade Agboola: Oh, no, it’s great weather. It’s sometimes a bit too hot.
74 00:13:40.510 ⇒ 00:13:44.060 Nikhil G: I was planning to come to Malta this, June, so, yeah.
75 00:13:45.220 ⇒ 00:13:48.820 Demilade Agboola: If you’re ever in Malta in June, just let me know. I should be around.
76 00:13:48.980 ⇒ 00:13:56.999 Nikhil G: Whatever the outcome is, We’d like to learn from your experiences, and then definitely keep in touch here.
77 00:13:57.470 ⇒ 00:13:59.950 Demilade Agboola: Sounds good, sounds good.
78 00:14:00.090 ⇒ 00:14:07.020 Demilade Agboola: Okay. So just, like, building on this scenario that we’ve…
79 00:14:07.240 ⇒ 00:14:08.869 Demilade Agboola: We’ve come up with so far.
80 00:14:09.850 ⇒ 00:14:13.280 Demilade Agboola: So in terms of, like, how we will build the final…
81 00:14:13.440 ⇒ 00:14:17.359 Demilade Agboola: reporting models. Would you want to use…
82 00:14:18.020 ⇒ 00:14:27.770 Demilade Agboola: how would you go about building it? Would you want to use a star schema or a normalized schema? And whichever you choose, why would you choose that?
83 00:14:28.250 ⇒ 00:14:34.170 Demilade Agboola: And when do you, like, choose a star schema when you choose a normalized schema? If ever.
84 00:14:35.930 ⇒ 00:14:37.509 Nikhil G: Yeah, yeah, definitely.
85 00:14:37.760 ⇒ 00:14:55.900 Nikhil G: So, yeah, it totally depends on what the depth, we want to achieve in the reporting style, you know? So, let’s suppose the star schema is nothing but, like, you have centralized fact table, and then the dimensional tables, you know, so that, like, you can…
86 00:14:55.900 ⇒ 00:15:09.250 Nikhil G: See, like, for which, let’s take example of the retail store, you know? Like, we want to see, like, how many orders and all that sort of stuff, you know? So, we can jump multiple things and get that thing.
87 00:15:09.310 ⇒ 00:15:16.770 Nikhil G: And it has the, extensibility as well, you know? So, it’s not built for a single thing.
88 00:15:16.770 ⇒ 00:15:35.059 Nikhil G: In the future, using that one, you can build multiple use cases out of it, you know? Rather, if you create the normalized view, that will be specifically only for that particular problem, you know, that we will be solving for, you know? So in the future, if there is any business change, we want to see another metrics.
89 00:15:35.060 ⇒ 00:15:54.819 Nikhil G: then there will be a lot of changes you will have to do on the single table. Rather, you… if you have multiple stars, like the tables in the star schema, then it’s portable. You can add new columns, you know, remove it, or change the formula, and still, like, get the results out of it. Like, if you want to build a new dashboard using this data.
90 00:15:54.820 ⇒ 00:16:00.850 Nikhil G: it becomes easier, you know? So, it depends, you know? Like, it takes time to build the star schema kind of data modeling.
91 00:16:00.850 ⇒ 00:16:04.240 Nikhil G: But it has, like, a lot of benefits in the long run.
92 00:16:04.530 ⇒ 00:16:07.579 Demilade Agboola: Okay. Okay, that’s fair, that’s fair.
93 00:16:07.730 ⇒ 00:16:10.570 Demilade Agboola: Alright, and then I think…
94 00:16:12.490 ⇒ 00:16:16.299 Demilade Agboola: Still within the same scenario. We have…
95 00:16:18.100 ⇒ 00:16:24.100 Demilade Agboola: A model within our infrastructure that, over time.
96 00:16:24.390 ⇒ 00:16:41.230 Demilade Agboola: has gotten really big, so now we have, like, 400 million rows in that table, and it’s starting to get very slow in our daily transformations. Assuming the worst case scenario, so, you know, just assume, like, this is the worst possible
97 00:16:41.230 ⇒ 00:16:47.179 Demilade Agboola: model that has been built. How would you go about debugging and making it efficient?
98 00:16:47.590 ⇒ 00:16:56.159 Nikhil G: Yeah, yeah, that’s a real world, you know, typical problem we face in, in the data engineering, or the…
99 00:16:56.160 ⇒ 00:17:08.679 Nikhil G: the data ecosystem overall. I have worked on the similar case, you know, rather than having, like, the millions of records, I have dealt with the billions of records in a table that scale, you know?
100 00:17:08.680 ⇒ 00:17:20.600 Nikhil G: So, as I mentioned, you know, like, the developer starts working in a way, you know, like, to make it easier to deliver faster, they build the, like, let’s say in the dbt, they create a viewer table.
101 00:17:20.660 ⇒ 00:17:23.470 Nikhil G: But as the data grows, company grows, you know.
102 00:17:23.470 ⇒ 00:17:24.560 Demilade Agboola: Yeah, definitely.
103 00:17:24.560 ⇒ 00:17:40.780 Nikhil G: you get a lot of volumes, and then it slows down everything, you know? So, the idea is to improve that thing is, like, definitely you can explore to use the incremental models, so rather than processing all the data every day altogether.
104 00:17:40.780 ⇒ 00:17:49.469 Nikhil G: you can do data processing every day, you know? So whatever new changes have come, whatever inserts, updates you want to target for, you should be only
105 00:17:49.680 ⇒ 00:18:09.630 Nikhil G: getting those, and then process that. So rather… so you don’t have to touch 400 million records every day. So rather, like, whatever, 5 millions every day you get, you insert. Again, like, with that, you need to consider what underlying database it is. Let’s say in the Snowflake, you need to have the cluster key, properly set in order to
106 00:18:09.640 ⇒ 00:18:18.560 Nikhil G: retrieve this data faster, and then reprocess it, and then store it back, you know? So, if you have used the wrong cluster key.
107 00:18:18.560 ⇒ 00:18:31.550 Nikhil G: then that hampers the performance as well, you know? So I have heavily work on the data scheming, the query plan, analysis, and then the optimizing the queries using the
108 00:18:31.920 ⇒ 00:18:38.609 Nikhil G: implementing cluster keys, and then the incremental, techniques in the dbt, yeah.
109 00:18:38.870 ⇒ 00:18:39.880 Demilade Agboola: Okay.
110 00:18:39.990 ⇒ 00:18:45.050 Demilade Agboola: Alright, that is very good.
111 00:18:45.690 ⇒ 00:18:56.300 Demilade Agboola: I think now, stepping away from all these scenarios, I have another question. So, if we have a client that wants a dashboard built.
112 00:18:56.580 ⇒ 00:19:04.659 Demilade Agboola: But they don’t give you clear, like, instructions on what metrics they want, or how they want, like, their dashboard to look like.
113 00:19:05.040 ⇒ 00:19:22.899 Demilade Agboola: How do you interact with such a client to get… to go from just, I want a dashboard, a sales dashboard, to very clear instructions that would allow you to be able to build out the models that you need and build out the infrastructure that you need, for this dashboard request?
114 00:19:23.650 ⇒ 00:19:24.930 Nikhil G: Yeah, yeah,
115 00:19:25.060 ⇒ 00:19:33.819 Nikhil G: I can give you one example from my experience, you know, how I dealt with this kind of requirements. Yeah, this is, like, typical again, you know, like, sometimes…
116 00:19:33.930 ⇒ 00:19:39.590 Nikhil G: And in that particular use, Sorry, use case, what happened?
117 00:19:39.870 ⇒ 00:19:45.070 Nikhil G: stakeholder came up with some dashboard, you know? They wanted to see metrics to…
118 00:19:45.380 ⇒ 00:19:48.129 Nikhil G: to make the decisions on. But still.
119 00:19:48.130 ⇒ 00:20:12.160 Nikhil G: they weren’t clear about the proper concrete requirements, what they want to see in the final reports, like, what metrics, what dimensions, and all that sort of stuff, you know? So, I understood that clearly, you know, because they don’t have clear directions on that one, but I tried to get as much information I can, you know, like, what problem they wanted to solve out of it, like, what are all the things they want as an input here, and then…
120 00:20:12.200 ⇒ 00:20:14.269 Nikhil G: What will be the outcome, you know?
121 00:20:14.270 ⇒ 00:20:14.870 Demilade Agboola: Okay.
122 00:20:14.870 ⇒ 00:20:35.140 Nikhil G: Based on that, you know, like, I tried to give the MVP one, minimum viable product, you know, so that, like, it’s a prototype, you don’t spend a lot of time to design it, like, you create one MVP as a product, and show it to the client, and then, based on that, like, you, then…
123 00:20:35.140 ⇒ 00:20:43.819 Nikhil G: goes into brainstorming, then they will open up about, like, what exactly they want, like, they will get more ideas about whatever I have provided, and then…
124 00:20:43.820 ⇒ 00:21:08.180 Nikhil G: based on that, like, they came up with a few more feedbacks, and then on top of that, also, like, I provided more MVPs, you know, like, improving that MVP on feedback, and then implementing the gaps and all that stuff, so that you can definitely go into the iterative phase, just deliver the MVP, get the feedback, improve it, and then
125 00:21:08.180 ⇒ 00:21:17.059 Nikhil G: Finally, decide on, like, what will be the final, the product will look like, and then you can decide on the acceptance criteria.
126 00:21:17.170 ⇒ 00:21:32.539 Nikhil G: And all that sort of stuff, cut over and all that sort of thing. Yeah, it’s typical sometimes, you know, like, whenever you don’t have the concrete requirements, you have to deliver in such a way that, so that, like, they can also understand the technical side of it.
127 00:21:32.780 ⇒ 00:21:49.590 Nikhil G: By not getting into technical, but still, they will understand the complexity, you know? It’s not that simple that, like, they will say, I want these numbers to be shown on this dashboard. It’s not sometimes straightforward to get from one column, like, you have to do join, or transform, and…
128 00:21:49.690 ⇒ 00:21:55.290 Nikhil G: A lot of things go behind it in order to show that particular thing on the front end.
129 00:21:55.870 ⇒ 00:21:57.600 Demilade Agboola: Okay, alright.
130 00:21:57.750 ⇒ 00:22:10.859 Demilade Agboola: And then this is my final question, and after that, we can also take questions from you, if you have any questions. So say a client wants a fast delivery, they want a fast turnaround on a request.
131 00:22:12.990 ⇒ 00:22:21.129 Demilade Agboola: But that means you would have to compromise the process that you normally go about building it. So, they won’t…
132 00:22:21.300 ⇒ 00:22:39.999 Demilade Agboola: numbers, the dashboard by tomorrow, right? And you know that, like, you would have to compromise on maybe architectural scalability, you might have to compromise on just even QA processes, something. Like, you’re going to have to make some certain compromises. How do you go about deciding
133 00:22:41.730 ⇒ 00:22:47.100 Demilade Agboola: What compromises to make, and how do you just handle that scenario, basically?
134 00:22:47.920 ⇒ 00:22:56.219 Nikhil G: Yeah, definitely. So, like, when we are dealing with some urgent matters, you know, like, you have to definitely consider some trade-offs.
135 00:22:56.590 ⇒ 00:23:13.709 Nikhil G: We definitely mark them down as a trade-off, you know, like, if you want that last board by tomorrow, then there has to be some trade-off on certain points, you know, the quality, you know, the actual confidence in the data, like, the accuracy and all that sort of stuff, you know?
136 00:23:14.050 ⇒ 00:23:18.709 Nikhil G: So you actually lay down all the points, you know, like, I can deliver this in one week.
137 00:23:18.820 ⇒ 00:23:37.319 Nikhil G: But you will have better quality, accuracy, confidence, and all that sort of stuff, but, like, I can provide you some tentative things by tomorrow, but then there will be a trade-off on the accuracies and all that sort of stuff. I will be still trying my 100% best, you know, but given the time, and then the…
138 00:23:37.320 ⇒ 00:23:52.119 Nikhil G: the urgency of this matter, I can still apply the same principles we have applied in the other dashboards, and still, like, we will follow the same rule, and then create this one, but I won’t be in position to…
139 00:23:52.490 ⇒ 00:24:08.159 Nikhil G: to do the 100%, the data validation QA and all that sort of stuff. Definitely, I can deliver that tomorrow, but after that, I will still continue that QA and then the data quality after… after that also, and then if I find any issues.
140 00:24:08.160 ⇒ 00:24:14.279 Nikhil G: we will rectify it and then provide you the, fresh numbers or whatever reports you want, you know? So it’s… it’s…
141 00:24:14.760 ⇒ 00:24:19.679 Nikhil G: Mostly about the communication, you know? Like, whatever the realistic things you can do, you know?
142 00:24:19.680 ⇒ 00:24:35.690 Nikhil G: It has to be, communicated. I would rather say over-communicated, so that, like, they, they also understand, like, what’s going behind the scene, so that, like, we are just not, pushing it back, for, for any,
143 00:24:36.010 ⇒ 00:24:44.279 Nikhil G: silly reasons, you know? So, yeah, definitely, like, once we start the processes, you know, like, once we get into the business, like, we…
144 00:24:44.370 ⇒ 00:24:58.539 Nikhil G: can come up with some pace, so that, like, even if some urgent matters comes up, we can take out some extra time as well, if there is any need, once in a while, not always, and then still deliver that kind of solutions,
145 00:24:58.760 ⇒ 00:25:01.419 Nikhil G: Considering the urgency and matter.
146 00:25:02.380 ⇒ 00:25:06.830 Demilade Agboola: Okay, alright, that’s fair. Thank you, thank you so much for your,
147 00:25:07.560 ⇒ 00:25:16.339 Demilade Agboola: answers to all the questions so far. At this point, it’s up to you if you have any questions or any things you’d like to know?
148 00:25:16.340 ⇒ 00:25:28.740 Nikhil G: Yeah, definitely. I was talking to Abeh in the last round, and it was really productive, you know, like, he shared how you guys are doing amazing work, you know, in the current
149 00:25:28.990 ⇒ 00:25:45.670 Nikhil G: this transformative phase, you know, like, more companies are working towards AI, and then, like, how they want to have the data foundation ready so that, like, they can use the AI to, you know, to actually get, meaningful things out of it, you know?
150 00:25:45.670 ⇒ 00:26:03.279 Nikhil G: So, just wanted to understand from your perspective, you know, like, what’s your day-to-day life looks like? How do you handle multiple clients at the same time, if you are assigned with the multiple clients, and then the tech stack, yeah? So, if you could just give me that overview, that would be helpful.
151 00:26:03.920 ⇒ 00:26:09.679 Demilade Agboola: Okay, yeah, I mean, there are a couple of ways in which that can work.
152 00:26:10.160 ⇒ 00:26:25.769 Demilade Agboola: So right now, I’m actually undergoing a transformation, in, like, my day-to-day. So, there are two ways I can answer that. I can answer that the previous way, and also, like, the current way. So I’ll start from the previous way. So, previously, I was, like, just an independent contractor working on multiple clients.
153 00:26:25.950 ⇒ 00:26:38.530 Demilade Agboola: So what that looked like was usually a bunch of, like, stand-ups, to be able to, like, sync with the internal team on what we need to do, and what’s going on, and how we need to get things across the line.
154 00:26:38.630 ⇒ 00:26:50.059 Demilade Agboola: And then I had time to, like, work on myself, on… work by myself, and work on the code I needed to do and push for the teams. So the models, in some cases,
155 00:26:50.850 ⇒ 00:27:10.060 Demilade Agboola: if things broke, I would reach out to me and let me know that, hey, these things have broken. So it was always, like, balancing. It was a balancing act. So some days you know, hey, we can’t do new things today because I’m going to spend time fixing old stuff, and I also need to go to another client after this, when I’m done with this work.
156 00:27:10.140 ⇒ 00:27:13.909 Demilade Agboola: So in that sort of scenario, it is kind of hard for you to…
157 00:27:13.910 ⇒ 00:27:32.740 Demilade Agboola: just do everything that comes your way, and it’s very important to communicate with the different stakeholders on the teams. So usually we have, like, roles that we call CSOs, so they’re client success owners, and they’re responsible for timelines and communicating if there’s going to be a push, a delay in the timeline.
158 00:27:32.900 ⇒ 00:27:51.590 Demilade Agboola: So usually it’s very important that you just communicate. You’re very clear about what you can do, what is possible, and what might need to be delayed. Nowadays, though, I’m trans… I’m trying… transforming my role. My role has been transformed to being, like, a service lead, so I’m responsible for all things data modeling.
159 00:27:51.730 ⇒ 00:28:00.499 Demilade Agboola: In this capacity, it’s less of a hands-on, day-to-day, building, but more of a thinking large.
160 00:28:00.790 ⇒ 00:28:11.499 Demilade Agboola: how do we go about building our strategy? How do we come up with playbooks? What are the standards for how we need to model? So sometimes I just need some quiet time to work.
161 00:28:11.500 ⇒ 00:28:21.429 Demilade Agboola: But also, I need to be responsible in touching base with anybody that is modeling, if they have any need for assistance, if there are any questions or concerns that they might have.
162 00:28:21.430 ⇒ 00:28:34.060 Demilade Agboola: Or just generally, just checking up on what they’re doing, and just to be sure that everybody that is modeling on different projects are moving things along. So, that’s kind of how that looks like my day-to-day.
163 00:28:34.760 ⇒ 00:28:49.010 Nikhil G: And, like, do you guys, like, spend a lot of times, meeting with the clients in the meeting, or, like, it’s us in communication over the email with Slack, and then you just get that work done? Like, how it is exactly currently?
164 00:28:49.420 ⇒ 00:29:00.199 Demilade Agboola: So usually, there are a couple of ways in which it works. So, we have our communication with us in Slack with the client, so we all have, like, external channels with the client.
165 00:29:00.330 ⇒ 00:29:17.549 Demilade Agboola: Now, while the client success owners, the CSO, are generally responsible for interacting and communicating with the client, part of the reason, or part of what we as, Brainforge want, are engineers that are also capable of talking to the client, right?
166 00:29:17.550 ⇒ 00:29:24.179 Demilade Agboola: So you don’t always… you’re not always blocked by a CSO. You can also say, hey, I have finished building out the models.
167 00:29:24.290 ⇒ 00:29:38.450 Demilade Agboola: can we have a call so we can go about the numbers, just to be sure that these numbers are in the ballpark of what you expect? Or these are numbers as accurate as you expect them to be? So you can also have that level of, like.
168 00:29:38.850 ⇒ 00:29:56.540 Demilade Agboola: scale. You don’t always have to say, hey, CSO needs to… no, you can also reach out to them and say, hey, can I book some time with you tomorrow at, you know, 1PM EST? And then they’re like, okay, yes, that’s fine, and you can have those kind of calls. But generally, we also try to have, once a… like.
169 00:29:56.580 ⇒ 00:30:06.510 Demilade Agboola: One time a week, we have a call with the clients, where we’re able to just tell them, like, this is the progress we’ve made so far. These are any blockers we’ve had, these are any things that, you know.
170 00:30:06.750 ⇒ 00:30:21.300 Demilade Agboola: happened over the past week. And we try our best to also just generally be in front of the clients, because one of the biggest issues we can… we have experience is not necessarily a lack of skill or technical issues, it’s just…
171 00:30:21.880 ⇒ 00:30:41.310 Demilade Agboola: lack of communication. So, if they feel like something should have been done, but something had happened that delayed it, but we didn’t communicate, then, yes, they would have issues, they would feel like, something is wrong somewhere. So, just being able to communicate and being ahead of things is very important, and…
172 00:30:41.640 ⇒ 00:30:48.740 Demilade Agboola: we do that through, like, you know, Zoom meetings, Slack channels, and… Also, just like…
173 00:30:49.030 ⇒ 00:31:06.769 Demilade Agboola: async calls, and async… sorry, async messages, and in some cases, we then use that to plan, like, impromptu calls, where we just hop on the call and just let them know, hey, this is what’s happening with this, this is what’s going on, we’ve done this, and we’re almost, over at the line with this as well.
174 00:31:07.310 ⇒ 00:31:16.740 Nikhil G: Okay, and if I may ask, like, what time zones do you guys usually work? Like, the time mixes, is there any flexibility around that one, or, like, how it goes?
175 00:31:17.230 ⇒ 00:31:20.080 Demilade Agboola: So we…
176 00:31:20.640 ⇒ 00:31:27.200 Demilade Agboola: Use the EST time, generally speaking, so that’s, like, Eastern Time, so that’s, like, New York time.
177 00:31:27.570 ⇒ 00:31:39.660 Demilade Agboola: The main thing is just, being able to communicate your hours, and being clear about your hours.
178 00:31:40.280 ⇒ 00:31:47.420 Demilade Agboola: Usually… like, for instance, I tend to work, like, the ESD hours directly, because I’m in Malta…
179 00:31:47.570 ⇒ 00:31:59.000 Demilade Agboola: For me, it ends up being, like, 2 to, like, 10, or 2 to 11, usually thereabout. So that works for me. But for some people, they…
180 00:32:01.640 ⇒ 00:32:20.400 Demilade Agboola: they end up working either prior to the EST hours, and then we advise that, like, most people work till 12 or 1 p.m. EST, and then you can log off. So you can start earlier, continue with the rest of the team, and then log off at, like, 1, or sometimes 2.
181 00:32:20.400 ⇒ 00:32:23.750 Demilade Agboola: But the idea is just, once you’re clear and consistent.
182 00:32:23.750 ⇒ 00:32:29.119 Demilade Agboola: And we ensure that, you know, you don’t miss meetings. So if, for instance, there I’m going to be meeting
183 00:32:29.120 ⇒ 00:32:39.670 Demilade Agboola: And you’ll be available, it’s very clear to communicate that. So we try our best to also try and have our meetings in the times that people are available for.
184 00:32:39.910 ⇒ 00:32:42.120 Demilade Agboola: And just communicate when…
185 00:32:42.690 ⇒ 00:32:51.819 Demilade Agboola: there will be a lack of availability, or, you know, you’ll be unavailable. So that’s kind of how we work. There aren’t any really, like, clear… we just… we focus on deliverables.
186 00:32:52.060 ⇒ 00:33:05.780 Demilade Agboola: Which is kind of also part of why, in our hiring process now, we’re, like, looking for people who are capable of working. Because, like, for instance, a junior engineer will need someone who
187 00:33:06.260 ⇒ 00:33:13.999 Demilade Agboola: is capable to, like, review their work, or just be there to give them direction. And such people, it’s hard for them to work.
188 00:33:14.340 ⇒ 00:33:33.950 Demilade Agboola: different time zones from the team. Yeah, because when you wake up, you’re just going to wake up to a bunch of questions, or things that you might need to, like, review. But someone who is, like, senior and capable can just actually solve things, push out the code, and the changes have been made. So, yeah, that’s kind of, like, what we’re looking for, and what we’re looking at.
189 00:33:35.520 ⇒ 00:33:36.700 Nikhil G: Okay, yeah, yeah.
190 00:33:36.880 ⇒ 00:33:45.010 Nikhil G: Sounds good. Interesting, exciting, so, yeah, definitely, I think we are at time, so, yeah, I won’t take my…
191 00:33:45.150 ⇒ 00:33:58.210 Nikhil G: any extra time of yours, but yeah, this was a really good discussion, really enjoyed, definitely excited to know more about the next processes, and, yeah, looking forward.
192 00:33:58.550 ⇒ 00:34:03.089 Demilade Agboola: Sounds good. I will give my feedback, and I’m sure Kayla will be in touch.
193 00:34:03.410 ⇒ 00:34:05.060 Nikhil G: Sure, yeah, thank you, thanks, Stephan.
194 00:34:05.060 ⇒ 00:34:06.910 Demilade Agboola: Take care. Have a good day.
195 00:34:06.910 ⇒ 00:34:07.969 Nikhil G: You too, bye.