Episode 8

Sadik Bakiu on the Future of Hiring: AI, Metrics, and the Changing Role of Data Professionals

TIM: What do you imagine will be happening over the next few years with data analysts, data scientists, data engineers? Like, how do you think that day to day role is going to change in a consultancy or in a company? Hello,

SADIK: years. I think the best we can do is predict in the next one or two years where everyone will be getting a lot of help from AI and then. Productivity will be increasing from all sides. And also it will be easier for people to jump in, into the data engineering, data analyst, data science world, and so on. However it's changing so fast and it's really unrealistic to make five year predictions, how this could look like.

TIM: Yeah, you're not Nostradamus. So it's too much of an ask for anyone to do that. Allowing you,

SADIK: as far as say, anyone that can tell you how it will look like in. The next five or 10 years, then they're probably lying to you.

TIM: or something like that?

SADIK: this time travel has happened.

TIM: and if you think about breaking down an analyst or a data scientist job, do you see certain things that are more likely to be used or more likely to be automated away with AI?

SADIK: I think data exploration, so figuring out what's in your data and so on, this can be this is very manual job, a lot of trial and error. AI can help a lot here and make this process much faster and by Point you the more interesting bits of the data or the more the outliers in your data so that you need to take care of them. And from a data analyst point of view, I think AI can give you recommendations of what are really some events that happened in the data that you have in front of you in order to focus on them and to handle them properly can be some, some fraudulent behavior. And then. You can see that it's not normal. It's not aligned with the normal pattern of usage. So these sort of things, I think, will get a lot of help from AI.

TIM: You also mentioned that it would almost like lower the barrier to entry to some of these roles. Is it going to be the case that we might have almost like domain experts coming from marketing or product who might not be data people at all, but they've suddenly got this tool that allows them to be like a kind of basic data person. Is that can you imagine that happening?

SADIK: Yeah, I think this can very well happen. I know these people that are domain experts have the know how probably can articulate in plain language what they want to achieve. And they only need one assistant sitting by their side, just writing the code for them or preparing the dashboard for them. With the direction how this is going with AI agents and AI everywhere. I think this will very realistically be possible.

TIM: Yeah. One thing I was chatting about today actually with one of my friends who's an engineer, was like the transition of an engineer from sitting there and writing code to maybe. Asking a large language model to write the code. And a lot of engineers would say that when they're writing their code, that's almost their thinking process. It's not like they start with the full design magically in their head. It's in the action that it emerges. Can you imagine us almost losing something then if we're not writing the code directly and we're asking them LLM, or will it just be like a slightly different skill and it'll still be the same process? Yeah.

SADIK: transition more from a writer to a supervisor. Of the, of what LLM is producing. So in order to be a good supervisor, you need to be an expert in it. I see there is some change or some paradigm shift happening, but yeah, I don't know how this will look like,

TIM: I feel like a lot of hiring is done in a gut feel and intuition based way. I'd say a lot of the different decisions along the way. Have you seen any approaches having like a more data driven way to hire people? And if so, how did they work?

SADIK: in order to have a data driven hiring You need to incentivize the right metrics. You cannot You need to be very careful what kind of metrics are you measuring so that you speed up speed up or optimize the process in the right direction. I think something very important to keep an eye on is how long does it take for the candidate. From the moment they submit the CV until they get the final decision. This is something companies need to take care of. Of course, these vanity metrics, like how many people apply for my job position or how many of the people we reject and so on, or I don't know, only 1 percent of the people get hired by us and so on. These are Nice for the feeling, but I don't think these are the right metrics that one should focus on when hiring. I think hiring should be as easy as possible for both sides, as little stressful as possible. An interview is never easy, but it should not be something stressful. And yeah, everyone should be considerate of the time of the candidate and also time of company.

TIM: yeah. I think the number of applicants is. going to be even more vain a metric than it has been in the past. If now people are using chat GPT or other models to apply on mass, suddenly there's going to be an explosion of applicants and it's going to mean nothing really, if 10, 000 people have applied for your job.

SADIK: this tells you nothing about the quality, the I don't know how interested are these people, how relevant are these people for your job position. It's basically empty.

TIM: exactly. We're hearing that increasingly, like from the conversations recently, that engagement seems a bit lower, like for each applicant, each person is involved. It's almost like the company then goes and tries to get them into the next stage of the process. There seems to be an increasing drop off feeling like, or companies are feeling like. Each candidate who applies is slightly less interested than normal, maybe because they're applying to so many roles or they're doing it automatically. They're motivation per application, if that was a metric is a fairly low at the moment, apparently. Yeah.

SADIK: Doing a shotgun application and see what what lands

TIM: The spray and pray approach. I think you call that sometimes. What about this in your experience a company is really hiring the best data talent, or are they just selecting those who looked best on paper? What have you seen in your experience?

SADIK: I've seen both. I think the biggest risk for companies is not hiring the right person, or as you mentioned, hiring someone that looks good on paper. And I think this happens to companies usually in the worst time possible when they are just getting started. So they don't know what a good fit would look like for the role or they have not figured it out yet. And they hire someone that would look, that looks good on paper. Does a nice interview, a nice chat, but sometimes lacks the depth. And then this this person would be very damaging for this data initiative that is growing in the company. Cause it's a lot of effort, a lot of investment from the from the very start. And probably the return on this investment is less than ideal.

TIM: Yeah, and I think if you're a small team or small company, each person you hire can Have a disproportionately big impact, either positive or negative on the culture and the way the team operates and everyone else in it.

SADIK: And in a small company, there is no place to hide.

TIM: What about this? I hear, particularly in the kind of talent acquisition and HR circles, a lot of focus on cultural fit. Something that I've always found a little bit vague in its definition. Like it's not necessarily measurable. Soft skills as well. Everyone would say they're super important, but maybe not super measurable. Like what is good or bad communication skills is subjective and debatable. Do you feel like companies ever overemphasize these in the hiring process and has that had any negative impacts in your experience?

SADIK: I think often, or probably not often, but very, it happens that companies focus a lot on cultural fit and this drags the hiring process and you might have one or two additional interviews to evaluate the cultural fit. This probably could be a bit shorter. In my opinion, if you can have a nice conversation with the person.

TIM: Okay.

SADIK: this. So yeah, a lot of companies are just doing this just so that probably it goes through the process. Probably they know something else that I don't know. But I think everyone should be considerate of the time. This is an investment from both sides and Interviews that don't have a lot of return should be skipped.

TIM: Yeah. A hundred percent. Yeah. It's amazing how long some processes are for really junior roles and a couple of people I've spoken to in the last week mentioned that as they've gotten more senior. The hiring process is shorter, not longer. And the last few jobs they've gotten was a coffee was just like a chat with someone and they got given the job despite the fact that they are on such a higher salary, like the importance of their role so much more yet the complexity of the process was zero. Isn't that weird?

SADIK: Yeah, it is. But, for these experienced people, they, I think they have paid it forward. So they have built their network or they have already shown their skill. And what what happens is, You just meet the right person at the right time and then it's it's done. But for junior roles, of course, you need to prove yourself and show that you are committed to this role. But in the end, a junior person does not really know how, to this role. Industry or corporate work. So you're not going to be able to evaluate much from that conversation.

TIM: for hiring most companies would have done manual CV screening until very recently. Now we're starting to see a lot of Chachapiti, Claude being used in the process, but until, I don't know, a year ago, it was like HR screen for a CV. Then interviews that were often. Not really a lot of stuff measured. It was more like getting a vibe check, a kind of gut feel of the candidates. Have you seen any ways that companies or consultancies have taken a more objective approach in the hiring process?

SADIK: so when one approach that we used to do, and we still do in my current company is having the engineers. Review the CVs because in the end, the engineers will be the people working with this person. They know what is more important for them and

TIM: share your

SADIK: How this how this CV fits to what they are looking for. HR is very important, of course, but HR is not the one that is going to be closely working. With the person that is applying, especially in the engineering department is if it's in the HR department. It's, of course, different story, but I feel having the CVS be being reviewed by engineers, it helps and gives a more realistic approach or a more realistic evaluation, how this. CV fits to what the company is looking for.

TIM: Yeah. I have I've quite a lot of empathy for talent teams because. It must be basically impossible to review CVs or applications of a job that you have basically no idea about, let's be honest. It would be like me trying to review CVs of lawyers, what the hell do I know? And so having the domain expert do it probably makes a lot more sense and it must be more accurate then. Yeah

SADIK: Writing some code or doing some engineering tasks, but I think it's very important that the engineers can choose their own peers. In the end, this is what makes their job easier or that makes their job more enjoyable. Hence I think this is a time well invested.

TIM: Okay. So they almost buy into it. into the process because they are the ones who benefit from the good colleague ultimately coming in.

SADIK: Absolutely.

TIM: there's different evaluation methods that companies use in the hiring process. There's quite a lot of academic literature out there measuring at scale, like how predictive each of these things are to this, like job simulation tests, IQ tests, personality tests. Unstructured interview, structured interviews, like a wide variety of tools. A lot of them seem to in the literature have basically no ability to predict performance. One of them is like years of experience is almost zero. Age is almost zero. So companies that still use a lot of these techniques that seem to have been disproven. Do you have any views of a why that is like why a company would keep doing something that apparently doesn't work.

SADIK: I think a lot is inertia. You have a system, you have a process and then You simply just follow the process and you never really look into the process. Is it really good? Is it bad? What, what's happening? And therefore it's important to have new people joining that can put some doubts into this process and take initiatives that re evaluate the processes. They might be perfect, they might be still working, but you need to be constantly re evaluating them and change the ones that are not working. Also it's important to, to be open to this change, of course. If everyone has tried to change a process in a company has seen that this is not very easy. However, you need to be persistent in this and eventually things will work out, but it's not the easiest task to do.

TIM: Yeah, for sure. It's everyone's. So change a verse, aren't they? No matter what it is or what it's about. And yeah, if you've been doing something the same way for 20 years, then probably you feel like there's no need to change.

SADIK: But world is not waiting for you, so you. Probably should be rethinking about it.

TIM: And what do you think about IQ tests in particular? Did you ever use those in hiring or yeah, what do you have used in IQ tests? Silence.

SADIK: if you take it twice in a row, you will Certainly have a higher IQ in the second time. I'm not sure if this is a realistic matrix, what what you get,

TIM: And what would you generally favor instead? What's a better measure for you in your view?

SADIK: for me, it's important to have a mix. You need a structured test probably some online test or some coding task, but you should also leave space for open ended questions. So things that show depth of knowledge from the candidate and the more strict or the more structured the interview process is, the less room you have to figure out these pocket of pockets of depth in the candidates knowledge. And these are in my opinion, very important to figure out because they show a lot about the commitment of the candidate, the knowledge And how dedicated are they to solve difficult problems? If you evaluate a fish by its ability to climb a tree, it will certainly fail. So you need to see where this candidate has his strong points and how deep this knowledge is and how deep. This his capabilities are

TIM: Yeah, and so by giving them an open ended problem, you allow them to show how they would solve it because they can just think of that themselves as opposed to sending them down a particular path.

SADIK: what usually happens in interviews with open ended question is the candidate gravitates to the part that they know best. And then you just need to follow along. Of course, you don't need to be an expert in all the domains. You should have an idea about how To solve this question and then just go along with the candidate and then poke for more details for more depth and challenge them so that you can also see how the candidate thinks how they would approach this problem how they have solved such similar problems before.

TIM: And as you say, it's such a trade off, isn't it, between having a really structured interview. Which then means that every candidate gets asked the same question, and it becomes easy to compare across them. But if it's too structured, it's almost rigid, and then you don't give a candidate an opportunity to really show the full depth or breadth of their knowledge.

SADIK: Absolutely. I think in life. You need to embrace change. Everything changes. If you commit to doing this particular job now, and then two years from now, you might be focused in a slightly different topic. Whoever has been in consulting, I think knows that. One project is doesn't look like the other and you need different set of skills. You need to be open to trying different things. And if in the interview you are basically testing for just one thing, you're not going to be able to find versatile people that can help you and the team in different directions.

TIM: Some hiring managers would dismiss technical skills in favor of soft skills. I hear this quite a lot and they'll basically say, Oh, I can teach someone technical skills. It's the soft skills that are hard to teach. You either got it or you don't. I'd rather hire someone with strong, soft skills and I can teach them the technical skills. Do you think they're being a little bit delusional is that I get the vibe that it's you're not a university, you need to hire someone with skills. You can't expect yourself to be spending all this time teaching people. Have you ever seen this approach backfire or even the opposite approach backfire?

SADIK: Yes, basically point is, if you hire someone with very good soft skills, but lacking technical skills, they are not going to be very productive in the team. And also the time commitment until you get them up to speed might be too high for certain companies. The opposite, of course, hiring someone with very good technical skills, but no soft skills. Also does not pay off very well. Often. It creates an hostile environment if you cannot really communicate with someone or if you, I know sometimes you need to give feedback, you need to change some things. And if it's very difficult to do this with one person or one colleague in the team, this affects the whole team. And it. Basically makes it not easy to work in the team. And now it's not the individual that is the problem anymore. So it needs, it certainly needs to be a balance. A company is is in the business of earning money, delivering solutions, consulting even more important because you are, Now working directly with a customer and if one team member is not up to speed either in soft skills or in tech skills, then this will be noticed by the client and has a risk on the reputation of the company and the client also affects the deadlines, the commitment and so on. And in consulting.

TIM: I'm not

SADIK: all about being reliable, taking something off your customers plate and making it easy for them. If you are making it harder to, to commit to the deadline or to stick to the deadline for them, this is counterproductive

TIM: And have you ever seen it happen where you've been in a consultancy, you've been on a project and one of the people staffed to that project has been dropped by the client where they've said, I just, I can't work with this person either because of their lack of soft skills or their lack of technical skills.

SADIK: more than once. It's it,

TIM: Okay.

SADIK: I've seen this multiple times and In consulting, it's a tough business and you need to convince the client immediately for, on your skills and Oh, they are. They're paying a margin on top for this so that they have experienced people or committed people to work in their project. And they don't have time to wait for you to bring up To speed your team.

TIM: Yeah they're paying, they're the customer, they need to get what they deserve basically. Yeah. Do you have a hiring hero, anyone that you can think back on who approached hiring in a unique or interesting way, or who you really learned a lot from when it comes to hiring data people?

SADIK: Yeah, it was my boss at data reply, Daniel Weingarten. He was awesome. And he'd basically come in the interview, live the spirit of everyone and then be very clear in what he's looking for. And he taught me a lot on what the good answer is, what the bad answer is, or what to look for at the candidate. So that You are not blindfolded by these people that are very good at soft skills, but now they're lacking the tech skills and so on. So you need to always be have a higher bar when hiring and keep everyone to that bar. So you don't need you should not lower your bar during the interview and so on. So a lot of such things I learned from Daniel.

TIM: Awesome. Great. Shout out for Daniel. And yeah, a few of those points really resonate keeping the hiring bar as high as possible. I feel like in my experience, when you're hiring. Cause it's so much effort. This is like a lot of interviews and this and that, and you're trying to do your day job. So it feels always tempting to me to go, this person seems good enough, even if there, you know they're weaker than your current team. And so I feel like there's always this temptation to finish the process and say, let's just hire this person. Like I can't do any more interviews. So it's good that he shared that insight with you to like press on. The most important thing is hiring the best person,

SADIK: Yeah, and it pays off so much more in the near and long term future. So it's really an important decision for everyone that is hiring.