Meeting Title: Brainforge AI Engineer Interview Date: 2026-03-09 Meeting participants: Kaela Gallagher, Marcus
WEBVTT
1 00:00:16.309 ⇒ 00:00:18.270 Kaela Gallagher: Hey, Marcus, good morning.
2 00:00:21.380 ⇒ 00:00:22.690 Marcus: Hey, hello, morning.
3 00:00:23.050 ⇒ 00:00:24.420 Marcus: How are you? Can you…
4 00:00:24.420 ⇒ 00:00:25.399 Kaela Gallagher: How are you?
5 00:00:30.020 ⇒ 00:00:31.290 Kaela Gallagher: Can you hear me okay?
6 00:00:31.790 ⇒ 00:00:33.160 Marcus: Yeah, I can, can you?
7 00:00:33.330 ⇒ 00:00:38.839 Kaela Gallagher: Okay, yes, just fine. Thanks for… for taking some time to meet with me, appreciate it.
8 00:00:40.600 ⇒ 00:00:58.060 Kaela Gallagher: Yeah, I guess just… so just to start off the conversation, I know you submitted, like, an application through our link, so thank you for doing that. I was unable to locate your LinkedIn. I’m curious, is your profile still active?
9 00:00:59.010 ⇒ 00:01:16.879 Marcus: Yeah, so my LinkedIn got blocked. The reason is, I hired a firm to basically apply on different roles, and I don’t know how they logged in or anything, so yeah, so that got blocked. I think it’s due to multiple login attempts, so yeah, I think I am currently in contact with the help team.
10 00:01:16.880 ⇒ 00:01:18.129 Kaela Gallagher: I guess when he came back.
11 00:01:18.710 ⇒ 00:01:19.879 Marcus: I don’t know what they’re doing.
12 00:01:20.520 ⇒ 00:01:30.139 Kaela Gallagher: Alrighty, no, no problem. Just like to have your profile in front of me, but I do have… have your resume, so I’ll refer to that.
13 00:01:30.500 ⇒ 00:01:36.329 Kaela Gallagher: Okay, yeah, I guess just starting off, I’m curious, like, what is putting you on the market for a new position?
14 00:01:37.290 ⇒ 00:01:46.710 Marcus: Sure thing. So, it’s been quite some time that I’ve been working with IntelliRent, and I worked on more than 7 projects there, so I think I had a huge learning growth
15 00:01:46.710 ⇒ 00:01:59.109 Marcus: learning curve there. Worked as an individual contributor, worked in a team of… worked across teams, also led two of the projects from the scratch. So I think, I worked on…
16 00:01:59.150 ⇒ 00:02:08.189 Marcus: approximately everything there that can be. So, yeah, the reason for me looking for a new role is that for past 3-4 months, it’s been…
17 00:02:08.190 ⇒ 00:02:13.719 Marcus: quite some linear kind of work that I’m doing there. There’s some bug fixes, some maintenance kind of stuff that
18 00:02:13.720 ⇒ 00:02:38.250 Marcus: required on the product. Everything that I worked on is complete and is on maintenance, or yeah, just bug fixes. There’s a very few features requested by the clients. Maybe, I would say, once in a blue moon. So yeah, that’s the kind of environment I do not want to be in right now. At this point of my career, I would say. I am looking for more dynamic, more career-oriented, more where I can see my technical growth.
19 00:02:38.300 ⇒ 00:02:48.269 Marcus: As I mentioned, I’ve been working as a lead, too. So, yeah, basically mentoring, is a part of what I can do. So, yeah, I think that’s what I’m looking for, and,
20 00:02:48.710 ⇒ 00:02:55.710 Marcus: to be very honest, I am not getting what I want, at this time of my career at Italy rent, so yeah, that’s the reason.
21 00:02:55.970 ⇒ 00:03:14.179 Kaela Gallagher: Okay, okay, got it. Well, at Brainforge right now, we’re hiring for a few positions, but all of them are fairly senior and would definitely have kind of, like, a leadership mentality that we’re looking for, and somebody that can kind of take a lot of ownership in what they do and work with a lot of agencies, so I think
22 00:03:14.180 ⇒ 00:03:23.539 Kaela Gallagher: You would find that in our positions. Right now, we’re looking for an AI engineer, an analytics engineer, and then a data engineer.
23 00:03:23.760 ⇒ 00:03:29.379 Kaela Gallagher: Which of those, three do you think you would be, like, best aligned with?
24 00:03:29.980 ⇒ 00:03:32.280 Marcus: So, I think it would be an AI engineer.
25 00:03:32.280 ⇒ 00:03:45.259 Kaela Gallagher: Okay, okay, perfect. Yeah, I think that’s what you had applied for as well, so that makes sense. Okay, what excites you, like, the most about working with AI?
26 00:03:46.350 ⇒ 00:03:57.550 Marcus: So, that’s a really good question. So, that’s what I asked when I was basically in university, that what I wanted to achieve, joining this particular
27 00:03:57.550 ⇒ 00:04:13.459 Marcus: industry. So, you know, there’s app dev, there’s web dev, there’s… then there was a very few machine learning at that time of… when I was joining, and then, there was a lot of AR, VR, or game where I wanted to be in, and the reason I chose
28 00:04:13.580 ⇒ 00:04:19.960 Marcus: starting with Python, the reason is I wanted to be a machine learning or AI engineer. The reason, again, is that
29 00:04:19.959 ⇒ 00:04:44.739 Marcus: I knew that there would be a huge boom in machine learning and AI. There would be very less development going forward. The reason, again, is AI that is… the LLMs that we have, everything is impromptu right now. A lot of LLMs, every company right now, I think most of the companies, I would say, are allowing developers to have some AI tools for the assistance of their development.
30 00:04:44.740 ⇒ 00:04:46.030 Marcus: and everything. So, yeah.
31 00:04:46.030 ⇒ 00:05:10.109 Marcus: Basically, that’s where I wanted to contribute, and that’s where I had the passion to be. So, everything, I think, if you go on Google and you can find the code for applications, you can find the code of web development, all of that stuff you can find there, but there’s a lot of brainstorming that’s required in machine learning and AI. The reason is you build from scratch, you have an idea, and you have to build it. So, not exact kind of stuff is present on,
32 00:05:10.490 ⇒ 00:05:20.490 Marcus: Google, for machine learning or AI. So, yeah, I think that’s where the most brainstorming was required, and yeah, I always wanted to be a part of that particular process, yeah.
33 00:05:20.670 ⇒ 00:05:35.580 Kaela Gallagher: Okay, okay, awesome. And can you tell me about, like, a time that you have built, like, an AI or ML solution that delivered, like, a lot of impact? I know you mentioned, like, that’s kind of the exciting part to you, is building from scratch.
34 00:05:36.300 ⇒ 00:06:00.449 Marcus: there are a lot of the projects, so currently, I don’t know if you guys work on AI agents, LLMs, or you guys work on building the models from scratch, so I’ll present both of the examples. So, the recent project that I worked on, you know, agents, LLMs are the buzzwords right now in the market. Everyone wants AI to be integrated in their platform. Either they want it or not, but they want it to be integrated.
35 00:06:00.450 ⇒ 00:06:03.130 Marcus: So, one of the projects that I worked on was
36 00:06:03.130 ⇒ 00:06:13.080 Marcus: Basically, the idea was mine. So, our company was hiring for a lot of people. They had… basically, I was connected with a firm that was hiring for a lot of people. They had a huge.
37 00:06:13.080 ⇒ 00:06:13.470 Kaela Gallagher: Job.
38 00:06:13.470 ⇒ 00:06:14.170 Marcus: a booth.
39 00:06:14.270 ⇒ 00:06:37.569 Marcus: posts. So they wanted the process to be automated in some way. The reason is, if you want to hire for 10 different roles, you would be having more than 500 candidates. How would you recruit, or how would you manage that kind of manforce? So, what I suggested, why not build an AI chatbot, or AI interviewer, we call it an AI interviewer, for the first round of interviews, so that the screening process is automated completely. So what we did is that what
40 00:06:37.570 ⇒ 00:07:00.580 Marcus: the questions that the recruiter would have asked, those was, those questions were fed to that particular AI that, then, whenever the candidate applied for that particular job, an AI interviewer link was sent directly to them. Those questions were asked directly by the AI and, answered by the candidate. We had, once they completed the interview, we had their audio, we had their video, we had the transcript.
41 00:07:00.580 ⇒ 00:07:24.029 Marcus: of that interview. So this way, then we had another agent of… which was basically, grading the transcript of that interview. So, based on our particular… our provided ground truths of the answers. So, yeah, basically, that’s where we, reduced the time of wedding a candidate from, let’s just say, 30 minutes to more, not more than 3, 5 to 5 minutes. So, that’s, what the process was.
42 00:07:24.040 ⇒ 00:07:41.119 Marcus: Yeah, this was completely led by me. I was the only engineer working on it, from the front end, back-end, AI integration, everything. And yeah, I got a lot of appreciation on this one, too. Other than that, I… if you talk about building custom models, so there was this one project where we had to basically
43 00:07:41.120 ⇒ 00:07:54.399 Marcus: scrape the data, from a website, which was a sports-related website. So we had to scrape a lot of unstructured data. Then we had to build a model, which was NER model, named Entity Recognition Model, which was to
44 00:07:54.640 ⇒ 00:08:05.259 Marcus: find out, the main keywords, we call them features, out of that unstructured data. And once we had those features, we basically stored them, we basically run different,
45 00:08:05.310 ⇒ 00:08:15.600 Marcus: algorithms on it, some were associated rule mining algorithms, and then there were some machine learning algorithms for some kind of predictions of strong and weak regions, then there were some other predictions, too. So, yep.
46 00:08:16.290 ⇒ 00:08:21.850 Kaela Gallagher: Okay, okay. I forgot to ask, but what kind of company is Intellerent?
47 00:08:22.080 ⇒ 00:08:24.970 Kaela Gallagher: Like, what do you guys do? Yeah.
48 00:08:25.550 ⇒ 00:08:29.979 Marcus: So, Intellivent is basically a platform where,
49 00:08:29.980 ⇒ 00:08:53.860 Marcus: it’s basically for a real estate agency just type of that. So, there are tenants, there are, yeah, if you want to sell your house, if you want to rent your house, so basically that’s kind of the platform that they have, but they do have multiple projects other than this, too. So this is the product that they have. This product is in the market for quite some time. It has a lot of user base. Now they work with different clients to, basically.
50 00:08:53.890 ⇒ 00:09:05.729 Marcus: have more in-depth in their software house, I would say. So, yeah, so we were basically outsourced to multiple… just like a consulting firm, we were outsourced to multiple other firms or other products, and then we were working on that.
51 00:09:06.300 ⇒ 00:09:19.180 Kaela Gallagher: Oh, got it. Okay, so it’s almost like, apartments.com or, like, a place where people can post their listings. Okay. Okay. Okay. Got it.
52 00:09:20.110 ⇒ 00:09:29.380 Kaela Gallagher: I’m curious… like… I’m curious about, a situation where you had an AI, like.
53 00:09:29.540 ⇒ 00:09:41.440 Kaela Gallagher: model or workflow, and it was, like, underperforming or, malfunctioning? Like, how do you kind of, like, diagnose and address issues like that?
54 00:09:42.380 ⇒ 00:09:59.010 Marcus: Sure thing. So, yeah, I think whenever you work on with AI, either agents or models, so there’s a lot of chance that, the model doesn’t do exactly what you require. So, either they hallucinate, or either they exactly underperform, or the model is… if you
55 00:09:59.010 ⇒ 00:10:14.469 Marcus: train on the data that is the only data that you are providing in the test cases. So, basically, that would be an overfair. So, yeah, a lot of cases occur. The reason usually is the data that you are providing. So, first of all, whenever you want a model to perform best.
56 00:10:14.470 ⇒ 00:10:37.249 Marcus: the first thing that I go to is the data. You have to make sure the data is correct, you have to make sure the data is consistent, you have to make sure the data is pre-processed or feature engineered correctly, the way that you want, the way that the model wants. So that’s the first thing that you have to look for, and that’s what I do. In 90% of the cases, that’s where the issue resides.
57 00:10:37.250 ⇒ 00:11:01.050 Marcus: you have to make sure that… in either case, sometimes you have terabytes of data, and the new data coming in exactly isn’t what your previous data was. So, the schema changed, or something changed in the data coming from the stream that you were scraping or getting it from. So, mostly the data, that’s the case when you have to make sure that everything’s consistent.
58 00:11:01.050 ⇒ 00:11:24.880 Marcus: if you… other than that, what we have to do is that there are multiple parameters that you have to weigh in, there are hyperparameters that you have to deal with, you have to tweak with to make sure that you are working with the best kind of, or optimized model. So, just like KNN, KNN, we have to basically get the value of K there. So, if you… if the data, or if the data is so much scattered that you
59 00:11:24.880 ⇒ 00:11:30.789 Marcus: that you… and you provided the value as 10, so it won’t perform. The reason is the data was
60 00:11:31.030 ⇒ 00:11:35.659 Marcus: A lot of… the data was basically scattered, and you won’t be able to find most…
61 00:11:35.660 ⇒ 00:11:58.799 Marcus: residing nearest neighbors. So, yeah, basically, that’s kind of parameters that you have to deal with, then there are different. So, in some cases, you do not use the exact model for your particular, or the right model, I would say, for your particular scenario. So, we have to make sure that the model that we are using is exact fit for our particular requirement, or the use case.
62 00:11:58.870 ⇒ 00:12:00.190 Marcus: Yeah, these are the things.
63 00:12:00.520 ⇒ 00:12:08.099 Kaela Gallagher: Yeah, I think that makes a lot of sense, and that’s something that, like, we really focus on as well, is, like, the data…
64 00:12:08.100 ⇒ 00:12:21.630 Kaela Gallagher: always comes first. Like, you can’t create an AI model or chatbot or anything like that, without, you know, a solid data structure. So, I think that aligns really well with, like, the motto of our team.
65 00:12:23.100 ⇒ 00:12:30.099 Kaela Gallagher: I know we just have a few minutes left. I wanted to see if you have any questions for me, or anything that I can help answer.
66 00:12:30.820 ⇒ 00:12:44.189 Marcus: Yeah, sure thing. A few questions. So, firstly, I would love to know a bit more about the role itself, what, what the guy coming in would be filling in, what’s team structure, how everything’s done in proven coach.
67 00:12:45.150 ⇒ 00:12:58.800 Kaela Gallagher: Yeah, so for AI team at Brainforge, it’s both client-facing and internal, so you would likely be supporting, probably about, like, 2 clients,
68 00:12:58.860 ⇒ 00:13:10.260 Kaela Gallagher: with… with their AI tooling, and we do ask that our engineers, you know, are capable of, like, communicating with clients, and joining client calls and being very, like, involved in that process.
69 00:13:10.340 ⇒ 00:13:16.750 Kaela Gallagher: But then our AI team also works on, like, an internal platform that we have.
70 00:13:16.750 ⇒ 00:13:38.949 Kaela Gallagher: So, for example, the platform contains, like, a repository of, like, our company brain, and, like, this Zoom call will be, like, added to that repository. And, you know, obviously it’s being recorded, and so, you know, I can go back and access notes, and,
71 00:13:38.950 ⇒ 00:13:44.270 Kaela Gallagher: have AI, you know, maybe take a look at it. So we have, like, a lot of different kind of capabilities.
72 00:13:44.270 ⇒ 00:13:53.639 Kaela Gallagher: on our internal, like, tool as well. So the AI team is, like, working on functions like that as well.
73 00:13:53.800 ⇒ 00:14:08.110 Kaela Gallagher: So, yeah, that’s kind of the overview of this position. The reason why it’s open is we’re just growing a lot right now. We’re about to sign, like, a couple additional deals, and we just need more support on the team.
74 00:14:08.950 ⇒ 00:14:13.449 Marcus: Got it. And how many engineers are there, and how does the hierarchy look like?
75 00:14:13.980 ⇒ 00:14:33.349 Kaela Gallagher: Yeah, so we’re a company of about 25 right now. We have a pretty, like, flat hierarchy. We do have, somebody that is, like, the technical lead for the AI team, so that’s who you would probably be partnering with the most, along with, like, your AI peers.
76 00:14:33.350 ⇒ 00:14:40.300 Kaela Gallagher: And I believe our AI engineers, I believe we have a team of about 4 or 5 right now.
77 00:14:41.410 ⇒ 00:14:42.090 Marcus: Got it.
78 00:14:42.500 ⇒ 00:14:46.610 Marcus: Yeah, my last question would be, what would be the next steps moving forward?
79 00:14:47.030 ⇒ 00:15:06.749 Kaela Gallagher: Yeah, so our interview process would be 3 steps from here. So we do a first round that is kind of like a culture fit, and just kind of understanding your experience in general. The second round is a bit more technical, like diving into your technical expertise and background.
80 00:15:06.800 ⇒ 00:15:26.590 Kaela Gallagher: And then our final round is a challenge. So we send you, basically, like, a GitHub link, with a practical challenge that you might come across, at our company, and so you would work on that for a couple hours ahead of time, and then come to the final interview with, like, your solution, and kind of.
81 00:15:26.590 ⇒ 00:15:34.380 Kaela Gallagher: discuss your solution to the challenge as… as your final round, and that would be a panel. There’d be a few people on that.
82 00:15:35.330 ⇒ 00:15:36.590 Marcus: Got it. Sounds good.
83 00:15:36.780 ⇒ 00:15:49.469 Kaela Gallagher: Yeah, cool. Awesome. Well, I will follow up, today with a link for the first round, and, just let me know if you have any questions throughout the process.
84 00:15:50.130 ⇒ 00:15:50.999 Marcus: Sure, I think, sounds good.
85 00:15:51.000 ⇒ 00:15:53.520 Kaela Gallagher: Okay, cool. Thanks for your time, Marcus.
86 00:15:54.110 ⇒ 00:15:55.749 Marcus: Yeah, thank you. Alrighty. Bye.