Episode 92

Aman Aneja on Redefining Data Leadership and Adapting Hiring Practices in the AI-Driven Hiring Era

TIM: We are live on the Objective Hiring Show with Aman. Thank you so much for joining us. It's really a pleasure to have you here. Welcome to the show.

AMAN: Thanks. Thanks, Tim. Pleasure is all mine. Thanks for having me and giving me the opportunity to have a discussion with you.

TIM: It's absolutely our pleasure. And I think a great place to start would be just a bit of an introduction about yourself. Just so our audience can start to build up a picture of who you are when we're listening to our conversation today.

AMAN: Yeah, sure. Sounds good. Hello everyone. I'm Aman Aneja. I am actually working as a director of analytics with a growth equity fund that's based out of London, and I've been working with them for almost two years now, and it has been quite a good learning experience for me. If you see my career progression or talk about my previous experience, I have been working with a single company before this. This is my second farm, right? So I have previously worked with a hospitality tech firm that's headquartered in India. I spent almost eight years with them, and I did quite different roles. And I went to different geographies as well. So this new job for me has been like almost two years now. It is a great learning experience for me, and it is still going on. Yeah.

TIM: Thank you for that summary and that intro. And yeah, I was looking and researching you a little bit before our call. And that was one thing that stuck out to me: longevity within one company, which is not that common these days. It's fair to say people often jump ship after a year or two. But what I found interesting is that you managed to continually, almost like, reinvent yourself in so many different roles and kept progressing through one company. I'd love to hear more about, like, how you did that and how the company supported that.

AMAN: Yeah, that's true. Correctly pointed out, yeah. I spent, like I said, eight years with Oyo. I joined at a very early stage, 2014, and the company started in 2013, if I'm not wrong. So I was like under 100 employees. with them. And I started as an operations associate. And eventually, I started to move in the space of analytics, and it was all on-the-job training, which I got. So it was all like exposure, which I got on the job. And the people, that's something that helped me a lot. So I was closely working with the leaders in the firm, and they continuously kept pushing me to try different things and learn different things. So from operations to analytics to data science to managing a team of 20 people and then eventually becoming a P&L leader. So that was the exposure that I got, and that's how I managed to evolve, in terms of my role and in terms of my learning. So yeah, that's what it has been.

TIM: That's awesome. And yeah, that's definitely one of the benefits of working in a high-growth. Aggressive company because it's like almost always new opportunities coming along, and they're all typically, I'd say, in my experience, quite meritocratic environments as well. So you obviously showed a lot of skill, a lot of quality in your work that then allowed you to get the opportunities that you then clearly took with both hands. And you now find yourself in quite a different type of business, but leading an analytics function. And I'm wondering if you could start to paint a bit of a picture around what you're doing day to day in your current role.

AMAN: Yeah, so As you said, right? So I'm currently working as director of analytics, leading data analytics and AI for the growth equity firm called Farview Equity Partners, right? So it's a small equity firm. I'm personally responsible for taking care of the whole data management and building AI models on top of it and also supporting the portfolio. Companies with required data analytics help whatever it is, like in terms of reporting, in terms of setting up structure, and in terms of building a pipeline. So from like when I joined Farview Equity, it was mainly like they were without. We had a backbone of data analytics, right? Everything was sitting out of, like, laptops and drives, on the SharePoint, right? I migrated everything to the cloud, and it's a completely cloud-based structure right now. We have our own. Cloud pipeline, which is there in Azure, and that's what I've been doing till now, and yeah, So this is how it has progressed till now.

TIM: Nice. And you mentioned also working with some of the portfolio businesses as well. Do you ever help them with hiring? Do you ever help them figure out who they should hire or how they should hire into their data and analytics functions?

AMAN: See, we mainly invest in the growth stage startups, right? Growth stage companies. Usually those companies do not have a data analytics function. At that stage, we are, we're talking about a company which is in a stage of becoming, let's say, 20 or 30 million firm in terms of the era, so at that time, they are not actually focusing on people or hiring of data analytics resources, so we Support them with the different kind of needs in, in the terms of setting structure for them in terms of like standardizing the segmentation of customer or any of the reporting. But as of now, the hiring support for them is nothing isn't not there for, from my side, at least

TIM: I wonder if that investment might start to get a bit sooner. Like in the stage of development of a company where they might start to be putting in data, especially for companies where it's going to be like a strategic value asset thing. Maybe that might be one of the first five or ten people in a company soon. Who knows? What do you reckon?

AMAN: Yeah, it varies. It depends on maybe the type of business we're talking about. So if it is a software product-based business, they will definitely be eager to hire more data analytics resources. Sooner. Where else, if it is service-based or part of a business, will you be more keen to have more, like salespeople and account managers, before you spend money on data analytics?

TIM: Yeah, that's probably the current mentality. I wonder also, based on what you've helped set up for some of these companies, are you getting in there yourself and giving them like a sort of minimal viable BI system that then becomes like self-service for them?

AMAN: Exactly. That's yeah. So that's the plan. And that's what I am currently working on. It's the project that's underway. But yeah, ad hoc requests and ad hoc support are always there. But in the long run, and that's what we are planning. So, like, whenever a company comes under our purview, like purview in terms of even before onboarding or before signing or before investing, we do the due diligence, right? So that's where we start doing the data work. And post-investment and post-onboarding them, the company becomes like part of the portfolio. Even then we support with the minimum requirements on the data sites. That's what we are doing right now.

TIM: When it comes to hiring, I'd love to have a discussion about the candidates and try to think of some kind of suggestions, some tips for candidates based on your experience with those candidates who maybe did really well and got hired or those candidates who bombed out and didn't get hired. If you think back to those you interviewed, are there typical areas where some candidates might fall down?

AMAN: Yeah, it's a tough one. It's like asking how one can crack all the interviews, right? I believe whenever a candidate is having a like interview. It's something that always comes across to me. If the candidate is trying to have a discussion instead of answering the questions only, that's where it becomes more interesting. And it is more like acceptable as interior from my side, right? So when you are having a discussion and then I think my, sorry, the candidate's key skill, apart from all the technical skills and apart from everything else, is like data storytelling. That's something that is a key attribute, and it should be there for any. Any data hire, right? You can always learn a new tool, a new skill, and in the current era of AI, right? So even if you have basic knowledge, you can put it in a different kind of Gen AI and can generate all kinds of codes, right? But it has to be more like, first of all, the skill in terms of the data storytelling. It should be there for any level of data hire. And then when you are having the conversation with the interviewer, it should be a discussion instead of just a question-and-answer round.

TIM: Is part of that almost like conversation flow down to the interviewer themselves as well in, like, setting that tone, do you think?

AMAN: Yeah, of course. It comes from both sides of the table, right? If the interviewer is just willing to have yes or no and just willing to know just like objective answers, then obviously you can't do much in that scenario.

TIM: I guess it must also partly be down to the crafting of the interview questions that are like open-ended enough that someone could answer them with their own mind, but not so open-ended that they're completely vague and they don't know where to start. Is there almost a certain skill to crafting those questions as well?

AMAN: Correct? Yeah. In terms of my knowledge or my belief when I'm doing an interview, I always try to put the discussion in a way that it's more around what the candidate is willing to do and willing to learn instead of just what they have done till now, right? So what are they open to having, right? It comes from my past journey as well. I won't say I knew everything when I was working with OYO, right? So it was like I learned everything, right? So my technical skill capability was almost close to null when I joined and when I entered the analytics space. But when I was working and I was learning, I was decent enough and doing my job. Fairly, so that's the platform, which I always try to give when I'm interviewing anyone. Try to learn what they're open to. And if they are like, That's one attribute or quality that I look for if they are adaptable to different scenarios, different situations, and if they are open to learning.

TIM: Is it fair to say that that trait, that adaptability, willingness, and ability to learn new things, is even more important? Now, if we're in this time of great technological change, where it feels like some technical skill sets are almost becoming redundant overnight and just having to learn all these new things all the time, is that now even more important?

AMAN: Yeah, absolutely. It is actually becoming more important while hiring and doing the team management in both processes, I would say. When you're hiring, you need to have these attributes for sure. Otherwise, it's becoming machine-only interactions, right? Which will eventually burst at some point in time. And it will not be good for anyone, I think.

TIM: And it also sounds like in your approach, you, as you said, are less fixated on their current skill set and more about what they could become. Which I can see the merit of clearly. But it also must be challenging because that's like a harder thing to evaluate, where someone could be in a year or two years as opposed to where they are now. So how do you unlock that in an interview? How do you begin to understand where they're going, what they're willing to learn, and what they're not willing to learn? Certain ways you try to probe them, certain kinds of red flags that might indicate a lack of a growth mindset or anything like that.

AMAN: Yeah, yeah, it depends on what scale or what level you are hiring, right? In the current situation, if I'm open to hiring any data hire, right? If I'm looking for any data resource, my discussion would be, And first of all, I will be open to it. I always believe this: it cannot be like a specific industry or field, which I'll be looking for. I'm always open to, like, different industries, different sectors. It all depends on the basic knowledge of tools, if I'm talking about the technology side of things, right? And then talking about, like, how I evaluate if they want to learn, if they want to brainstorm at least. Even if they are not able to do it properly, which is obvious, if I'm giving him or her a scenario from our business, they won't know exactly what they should be answering. But how much effort they are trying to put in is something that gives you the level of. Level of what do you say? Adaptability, which they want to put in, right? So that's what it is all about. Will you, when you're interviewing, give different scenarios? It could be from your industry. It could be a very basic thing. It could be a basic estimate as well, right? If their approach is good, if they are trying to solve it, if they are trying to put in an effort, that's where I believe the person would do. good In the coming future.

TIM: And so a good effort for you in the interview would be like they're thinking about it closely, that kind of asking follow-up questions, they're engaged, they're not just giving you the first thing off the top of their head; is that part of where they would demonstrate their interest?

AMAN: Exactly. Exactly. Obviously, it all starts when you take the time to think through, and then you structure it in a way that, like, you put it, What was the objective, and how did you come through to the process of solving this and that? The structure of a whole discussion is what gives you confidence, right? That the person is actually understanding and trying to actually get through the solution. So the whole structure, like what is objective, what are different things that can impact that particular problem statement, what is the hypothesis, and the whole structure, if that is there, then I believe Even if the answer is not correct, the approach is correct, so that should do.

TIM: In that kind of adaptability, willingness, and ability to learn. It's part of that wrapped up in being able to take feedback as well. Do you ever try to, let's say, when they're talking through their kind of solution to these scenarios you've painted out? Do you ever dig in a little bit and criticize them and see how they respond to your, let's say, constructive criticisms?

AMAN: Not exactly criticizing them, but yeah, obviously I try to divert a bit in terms of giving some additional, for example, I'm giving a rough, very basic estimate of how many people like traveling through, let's say, the London underground train in a particular day. It's very basic, right? Let them think it through, let them, yeah. Structure it. And then I always say, take your time and then structure it and then come back with the answer. Take your five minutes, right? And when they are going through it, and when I'm feeling they're doing good, I try to take them off track. What if there is a signal failure? What if there is an alternative option available that is cheaper? And all this just to understand when anyone is working, like it helps you understand, like, in the long run, when the person will be working on any project and when management says our objectives. We are no longer focusing on this thing. We have to focus on this. So how adaptable are they for the different situations in different scenarios, right? So that's what I try to understand by taking them off track slightly.

TIM: Literally off track, in this case, if it's a question about train market sizing, are there any? If you think back to the different candidates who've failed, either at the interview stage or beforehand or another stage, are there really recurring common problems that would lead to these candidates failing?

AMAN: It depends. But I think if I look back and try to categorize, I think it's majorly around data storytelling. That's what is missing on the technical people. Because they always keep on focusing. I'm not talking about the current scenario. It might change because people's focus will actually change now because coding is no longer a big task or a big skill, right? But it used to be only focused on technical people. I'm talking about, let's say, higher-up data scientists or data engineers, right? They always miss this skill, which eventually leads to stagnation off their position in the next stage. I don't feel like they will be able to excel that much if they don't have the skill, and that's where I always like that. That's one key thing that I have noticed in many candidates and has resulted in, let's say, maybe not taking them as a hire.

TIM: I wonder what you think about this, as you mentioned with LLMs getting to a point where maybe they could do all of the coding for us. You might have a ludicrous idea to try to manually program, and it's just going to be a given that you're going to prompt an LLM; it's going to do it for you. Is it something to then be said for the kind of mix of skills that a data person needs to have is becoming less technical? Arguably, the data storytelling might be even more relatively important than it has been. What do you think about how that blend of skills is going to change?

AMAN: Yep, so I think, yeah, it does become, it is changing rapidly. It's quite good to adapt to the technology and the new technology, which is LLM, and it is always good to hire maybe a fresher out of college or with less experience, which will eventually cost you less, right? They don't need to be trained in terms of the coding. They need not understand, like, end to end, maybe any language, Python or anything, right? If they know the basics, they can code everything. Yes, the focus actually changes on the level of higher, which you are doing, and then I think, yeah, it is actually changing even in the current scenario I'm looking at in the market.

TIM: And is it a case then that. I guess some people might have to have a bit of a rethink of their profile and almost like their self, because I don't know about you, but I feel like for some roles, particularly, let's just say coding, some software engineers would describe themselves as coders. Like, it's almost tied up in their own identity of who they are, is the fact that they are coding and programming. But if that is going to be like disappeared in the next year. underneath our feet. Will people in these roles have to just rethink who they fundamentally are, like what they do, because it's not going to be the thing they've been doing for maybe 10 or 20 years?

AMAN: I don't think so. It will go away completely. But yeah, the level of skill set, you won't be needing like high-experience people, even if, and like I said, even if you have the basic experience, that will also do. You don't need to hire, let's say, a 20-year experienced person for doing coding because he or she has done this so many times; even if you have done two or three years, you know the basics. Then you have the knowledge of how to code. It could be software engineering; it could be data scientists. It could be any other field where you have to, like, code. I think it'll eventually become a change in the world that you look out for a person with lesser experience, which will actually cost you less.

TIM: One thing we're hearing a lot of, particularly in the U.S. and European markets, is companies that are getting inundated with many applications through LinkedIn and through other job portals, especially for technical roles like data scientists. And part of the feedback I'm getting, which is a bit anecdotal, but still, is that. They're getting a lot of applications that are starting to look a lot like each other and sound like each other and that look amazing, as in they look like a very good match for the job. What I assume is happening is people are using ChatGPT or Claude or what have you to optimize their CV for the job ad, which then makes me think this is going to be tricky. Do we need something other than a CV to screen on? Is this almost some new data set we need? Because if we're getting a thousand applications and 500 of them look amazing, how are we going to choose among the 500?

AMAN: Mean, yeah, this is a problem statement, which should be there for many recruiters right now. People using OpenAI or any other NLM, as this is a free service, right? So everyone uses it for generating the CV or the cover letters, right? I believe using AI for different other processes can help us in identifying these candidates or at least having a preliminary assessment, something that I guess Luba also does. So that's something; it can help you actually in identifying the correct set of people. It could be a simple quiz or it could be a simple coding test that cannot be done by AI. You have all the tests and checks and balances that can be done on the test, right? So if you are on the screen, you can't switch the screen; you can't use any other. You have to be looking at the screen. So all those things can be done. And I can monitor it. So I think this can help in actually filtering out that kind of lengthy, you can say, many candidates, which are just. Copy-pasting the whole content and just applying, right? So it will be useful if AI is used end to end rather than only on a particular set of processes.

TIM: Yeah, I feel like it's going to have to be something like that, like a skills-based screening approach. Because at the moment, I think part of the challenge with screening is we just don't have a good data set. To decide whether or not to bring the candidate in, because it's even before AI, ultimately, it's a document written by someone about themselves, like, it could not be more biased. And that's just from that perspective. Then on the other side, we've always had the issue of the person doing the screening and all the biases they bring to the table in terms of making a selection. So I feel like the CV screening step was always hampered. And now it's just even more so because. It's just this mass generated, almost like spam. from an AI. So we're going to need to unlock some new data sets, I think, and an evaluation of their skills is a better one, I think. Certainly from what I've seen, speaking of biases, are there any things we should be wary of if we're starting to add? AI to the mix, any kind of new biases that might introduce or any other concerns that are legitimate there, do you think?

AMAN: I think when it is worry as well. If the machine is doing everything, the whole hiring process, I always believe this; even if you are targeting, let's say, to automate the whole hiring process and using AI, there has to be at least one or two rounds of personnel. Touch base with the candidate. It could be like a cultural round, or it could be a manager round, or any such kind of discussion, right? Which can help you do the correct hiring rather than just. Leaving everything on AI, right? So that's what I believe you can do in terms of removing the biases. But in like removing biases within the AI, I'm not quite sure as of now.

TIM: How would you feel personally right now if you applied to a hiring process and it was pure AI, like you got it from start to finish, from applying through to offer, and you never interacted with a human? Would that be cool or scary, or what would your reaction be?

AMAN: I think it would be exciting for sure, like in a way that it's a new experience. I personally won't be comfortable with the whole process. If I'm doing end-to-end. The thing with the machine is that there is no person or human angle attached to it. I won't be comfortable; that's what I believe.

TIM: And that lack of comfort would come from the fact that it's risky. Cause you'd be joining a company when you haven't met anyone there. Is that part of the discomfort, do you think?

AMAN: Yeah, of course, that's one of the reasons, and also, like, I don't know who the person I will be working with is, right? That's like, at the end of the day, I like what I believe is the person or the people you're working with within the organization that impacts your journey. And it's my personal opinion that the whole journey at OYO was. All dependent on the person or people I was working with, right? So that thing impacts a lot. And if it is missing, then personally, I won't be comfortable at all. Even if they are, like, paying me millions, I won't be comfortable going with them.

TIM: Yes. I'm thinking back to the last job I applied for and got, which was 10 years ago. But I'm just thinking back to the hiring process there. And of the people that I met, I only met two people. One of them was a co-founder who was no longer involved, and the other one was a co-founder. They're the people who hired me as one person. All the other people I've met at the company, some of my best friends, I never met in the hiring process at all. Didn't know who they were until I started working there. If I take my own personal experience If I went from one person to zero, it doesn't seem like a huge difference in some ways, like in terms of the risk, but what I imagine will happen with AI is we start to implement it from the top of the funnel down. So as you say, maybe the screening step with the CVS first and then a skills assessment, and then maybe as an interview assistant with a human involved. But ultimately getting closer and closer to the end decision, I suspect, just because that's the most important decision. So you probably leave it up to the smartest human you have, probably.

AMAN: Yeah, exactly.

TIM: We'll see how that plays out. One thing we often experience in hiring, probably any project, is if you think about it, there's like the speed, cost, and quality. That kind of triangle of trade-offs, and it's hard to have something really fast that's high quality and cheap. You've got a trade-off on one of them. For hiring, it's often about time to hire, so trying to hire as quickly as possible. How do you think about this kind of trade-off between hiring quickly but also making sure that you don't make a bad hire?

AMAN: I've been in the situation in the past, not right now, but I've been in the situation in the past. It used to be this pressure hanging over the head, right? To hire the person as soon as possible because of x, y, z reasons, right? So a person is leaving, or we need a person because we are starting a new line of business or whatever it is, right? But I believe even if it takes, let's say, a few weeks extra. To hire and finalize the candidate, it's absolutely fine to do that because if you hire the wrong person, wrong in terms of not being appropriate for the job, then it will cost you exactly twice or even more than that, right? to it, rehire or replace the person? So it's always good to be comfortable in terms of what and quality and if the person is appropriate for the job. So I tend to take time in terms of I do not hurry; let's close the person in, let's say, two or three days.

TIM: Yeah, I think that's the right trade-off to make. Because, as you say, the cost of the well, or what's sometimes called a bad hire, as you rightly point out, the bad person, or the wrong hire, or the regretted hire, or whatever you call it, is just catastrophic because it's not just the cost of the rehire and going through the process and what have you; it's also just like the mental strain of everyone involved.

AMAN: Exactly.

TIM: Someone's losing their job; you have to performance manage someone. Oh my God, it's just awful. I don't know how you quantify that, but it's, yeah. wants to go through that.

AMAN: Yeah, it impacts a lot. Like the whole team's motivation also goes down. And yeah, try to keep it relaxed rather than, like, based.

TIM: One thing I'm always struck by is, no matter how objective and data-driven we make the process, no matter how much we collect. Hiring accuracy is going to be probably no better than 70 or 80 percent anyway, even with the best information. Like, I always love hearing stories from football clubs and sports where they've just spent, like, a hundred million pounds on a player. It's like the cost of an entire business's salaries for several years. They've just spent that on one player, and the player gets into the first training session, and his teammates are like, This guy is crap. It's dead. 100 million pounds. So if a football club can waste that with all the data in the world, then of course, not every hiring decision on a data analyst is going to be that perfectly accurate either. So it's good to bring it back to that kind of level of humility and expectation to not overthink how good we can be at hiring.

AMAN: Yeah, that's right.

TIM: What about, yeah? Bigger picture. I sometimes think of eyes, approaching magic at the moment. But if you had a magic one, be it AI or otherwise, how would you be fixing the hiring process? How would you change it? What would some future perfect-looking hiring processes look like to you?

AMAN: Yeah, we talked about this in bits and pieces in our discussion, but I believe it's like it's good to adopt the technology as much as you can while doing your work, but I believe that has to have the human angle, which will never be able to be replaced by AI. I say ETC, which is like emotions, trust, and confidence. These three things, I don't think AI will be able to understand. About any candidate or the interviewer, right? So these three things need to be done by a person, right? So as much as automation you want to do, you need to keep the human angle in the hiring process. And as far as I know, no AI tool can actually tell you all these three things as of now, right? So yeah, you can go for the whole process automation, but you still need to have a couple of discussions with the candidate in person or telephonically or whatever it is.

TIM: Is or would part of the challenge in even contemplating automating those three things be that they're quite subjective, and that even you could have three different interviewers, and they might evaluate a candidate along three different ways, according to their communication, their trust, and their emotion? And so is that, yeah, is that subjectivity, maybe? Is that subjectivity part of the challenge that we're evaluating people along lines that even we can't put our finger on perfectly, how to measure them?

AMAN: Correct. That's the point.

TIM: And how do we resolve that? Is that just what it is? That is what it is to be human. And that's just a principle, and it's never going to change. And that's

AMAN: That's what I'm trying to say. It needs to be the same way. These things should not be replaced by the machine. Otherwise it will be like,

TIM: terminated.

AMAN: We don't know; we don't need people anymore.

TIM: Yeah, and that's one interesting thought experiment as well, isn't it? Before we were talking about that kind of mix of skills, maybe getting a bit more relatively softer in a data candidate. The only other thought I've had is maybe it's the exact opposite, which is now if your average colleague in a couple of years is going to be an A. I. Not a human. Maybe you really have A.I. communication skills. That's the most important thing as opposed to human. Maybe that's the route we're going to go down.

AMAN: That might be a key requirement of a role; that could be the case.

TIM: If you could ask our next guest one question about hiring, what would that question be?

AMAN: I think I would like to understand the same thing that I've already mentioned. Considering the AI and its adoption increasing rapidly, would you be willing to hire a fresher from college rather than hiring a person with 10, 12, or 15 years of experience? Would you gun for maybe fresher junior resources more? Then spending a fortune on senior roles, right? So that's something I would really like to understand.

TIM: Yes, we will level that at one of our guests in the next week or so and see what they say. I'm interested to hear their thoughts as well. It's been a great conversation today. We've covered a lot of different ground. Thank you so much for joining us and sharing your insights with the audience.

AMAN: Sounds good. Thank you so much, Tim. Thank you for having me. Have a good one. Bye bye.