Episode 72

Mrunal Tipari on Balancing Data-Driven and Gut Feel in the Hiring Process

TIM: We are live with Mrunal on the Objective Hiring Show. Thank you so much for joining us.

MRUNAL: for having me

TIM: It is our sincere pleasure, and I would love to kick off our conversation today with a bit of an observation that I've made over the past few years, and that is that there are a lot of companies out there that will claim to be very data driven, very informed by data, and a lot of data leaders themselves who would be in their day jobs either running product analytics or marketing analytics or sales analytics. and we'll be preaching the value of data, but then when it comes to hiring, we almost take off the data cap and put on another cap, the more kind of intuitive gut-feel-based cap, and I'm wondering if you have any thoughts on why that is the case and why we move towards this kind of intuitive hiring decision process rather than taking a data-driven one.

MRUNAL: I think the reason it's more about gut feeling when it comes to hiring is because certain qualities can just not be measured. It's hard to measure things like a person's ability to learn or curiosity or, as simple as just communicating with the teammates. So this kind of quality doesn't show up on a CV, and it's not something like we can measure through a technical test. Yeah, I think that's why it's more of a gut feeling, but again I'm not saying it should be totally biased to gut feeling as well. I think that should be a right balance between where you can use the data and where you can use the gut feeling.

TIM: Yeah, and I feel like for a lot of these, let's say, more subjective elements in hiring, let's say not the technical skills, which are a bit more objective, but maybe a combination of the softer skills and the, let's say, cultural fit most companies would call it, they are inherently a bit more subjective in that there's not necessarily a right or wrong way to communicate. But I still tend to find that a lot of hiring managers would take on a spectrum of gut feel to data-driven, a very gut-feel-based approach where maybe there's some opportunities to use data to make it a bit more objective even if it is inherently subjective. What do you think? Yeah, I think we can use data now that you've said that, so, for example, if they have an experience where they have given, let's say, presentations or just like being on a podcast as well, that could be a way of telling, Okay, they have some soft skills or subjective skills. Yeah, I think that kind of assessment can still be data-driven and can be used as a data-driven site. Yeah, I'm interested to see where AI will take this, actually, because I feel like maybe one of the benefits of large language models is they're good at dealing with unstructured data, and so maybe in the future there'll be a process where you could get a more holistic view of a candidate. look at maybe the GitHub profile podcast they've done, a Medium blog article they've written If there's some kind of way to automatically aggregate or grade that, maybe that will be something that happens in the next few years.

MRUNAL: It depends on what parameters the model has been trained on, so a couple of years ago one of my colleagues gave me a book called Weapons of Mad Destruction. I'm not sure if you read that, but it had a lot of examples and different industries, like how the models have affected the outcomes. So if we are talking about hiring, it was like the automated system unintentionally introduced biases and filtered out great candidates, so because they had seen in the past that only people from certain colleges were great candidates, they were just picking those colleges, for example, or men and women, because men were, for example, very high, and then that's why it was just filtering out women. So I think the parameters considered in that model are very important, so carefully designing the system is the takeaway.

TIM: Yeah, for sure we don't want to unleash some future with AI ruling everything. I feel like, though, in hiring in particular, the current scenario, the current way hiring is done, is, in my view, so biased that it's hard to imagine how introducing AI would make it worse, so I'll give you an interesting example. There have been a lot of experiments over the past 10 years in different countries. I know there have been these experiments in Australia, America, and England where the researchers will apply to different roles with different companies with CVs, with the only real difference among the sets of CVs being the name that's on the CV. So there's one in Australia. I think in 2022, Sydney University had several thousand CVs in three groups: one group had Anglo first and last names, the second group had Anglo first names and Chinese last names, and the third group had Chinese first and last names. Otherwise, the sets of CVs were pretty similar, and there were thousands of them. They then applied to tens of thousands of jobs around Sydney and Melbourne and measured the rate at which the applications got a callback, and so for the first group, which had the white first and last name, they've got a 12 percent callback rate. The last group, which had the Chinese first and last name, got a 4 percent callback rate. If you apply for a job in Australia with a Chinese name, you have one third the chance for a callback compared to all else being equal, which is appalling, let's be honest, and so I feel like hiring is one of those things that's already so biased by humans that it would have to be a pretty bad AI system in my view. to make it definitely, yeah, I think it shouldn't be based on any kind of demographics or the person's personality or whatever it is that defines that person but just based on the skills and the qualities that person has for, let's say, hiring managers and companies who are taking a very intuitive approach to hiring, so let's say their idea of hiring is, All right, let's do a quick coffee chat with someone and just get a feel for them, do like a vibe check, and everything's very subjective, and they just get a sense of whether or not they like the person or not. If that's their hiring approach, are there benefits to them sitting down and saying, Actually, what can we measure in this process? we decide we want communication skills Maybe we could start to define what that is and measure it. Would there be some benefits to moving at least a little bit towards the data-driven end of the spectrum, do you think?

MRUNAL: Yeah, definitely, because it helps with the initial screening process. If there are thousands of applicants, you're not going to sit with each and every one of them and have a coffee, right? It's not financially feasible as well, and also everyone's time is valuable, so I think the initial screening process is where the data-driven approach can help. And then if you want, if you like that person, then you can have in-person interviews because, as we have talked about as well in our introductory call, soft skills are equally important to technical skills.

TIM: And speaking of those soft skills, like how do you personally make that trade-off between the technical and soft skills? You're looking for a combination of both. If you have to choose one, which one would you choose? I'd love to hear your thoughts on that.

MRUNAL: Great question, as we are a startup, and the members of our analytics team have to wear multiple hats, working as a data analyst one day and then maybe on a data science or engineering project the next, so this kind of adaptability, I think, needs a balance between technical expertise and communication skills so that the team is aligned. But again, I'm not saying that it should be a 50-50 percent split, no, not at all, but at least like 10 to 20 percent of the candidates overall profile, at least they should be able to talk with their teammates or be proactive and ask questions, because I cannot stress enough how important asking questions is. If you don't ask questions, you're not going to grow, especially if you're a new hire trying to catch up quickly in an environment, especially if it's a startup, which is fast-paced and more like a collaborative environment. Asking questions is really important; that's the first thing I always tell my team or the new hires: there are no silly questions, so, yeah, it should be a balance.

TIM: And what about startups versus large organizations? Then are there any aspects or any bits of the data professional that are going to be a little bit different between those two organizations or anything different that you look for?

MRUNAL: Yeah, so I feel that having worked in both MNCs and now in a startup, I see that in startups you are looking for generalists, as we talked about, wearing multiple hats, and then people who can learn quickly or adapt to evolving needs because we often hire recent graduates. They are a great fit because they are still exploring their career paths and can discover that, like, whether they prefer a data analyst when they're working on an analyst project or science or engineering. So I feel that at TV AdSync, at least I've seen this kind of flexibility of recent graduates as a strong advantage, but when we talk about the larger organizations, I think they look for specialists, but they like a person, a data engineer with this technical stack and so on. If that person comes into a startup, I don't think that person will like working because he or she has to work on different kinds of projects as well.

TIM: I guess maybe also part of the clash might be just the rate of change; like in a large organization, typically, on average, they operate a little bit slower, a little bit more methodical; everything's a bit more planned out, whereas a startup has that level of chaos that certainly isn't going to be everyone's cup of tea. I wouldn't have thought

MRUNAL: True, yeah, you have touched exactly on the right point; that's definitely true, and some people are just built for startups. I feel some people like to innovate and always try to think out of the box and bring in different perspectives. For example, as you said, it is slow in the larger organizations, but in startups, if, for example, we are doing a process in a certain way but a new person comes in and says, Okay, we actually can do it in a better way, so that transition or change of the process is very quick and efficient.

TIM: And in your hiring process, how do you currently try to coax out those people who you feel like could be a good fit in that startup kind of environment as opposed to ones who might struggle to adapt?

MRUNAL: good question I think by just talking so, after initial CV screening, when the person comes in for an interview, we just talk about whether they have done anything outside their curriculum that just shows that the person has the will to grow or learn new things because even if we are hiring a recent graduate, Where they are very good at technical and soft skills, there's always a missing part, which is getting to know the industry. We are an ad tech company, so that will always be a missing piece in that candidate's profile, so if we see that person is okay, has the ability to learn, we know, okay, that person is going to be okay in the company. Yeah, so I think from just talking and seeing what they have done and how they approach things and how they solve problems, I feel it's not necessarily that if some if a question is asked, it's not necessarily that they have to solve it, but just the approach tells more about the person.

TIM: I feel like hiring is one of those things where the more transparent we are on both sides as soon as possible, the better it is for everyone, because the kind of environment you offer is going to be amazing for some candidates and probably dreadful for others. It would either be your dream job or your worst nightmare. and the sooner we have these conversations in the process and the more forthright and transparent we are, the more I feel like it just saves everyone a lot of time.

MRUNAL: No, definitely setting expectations is one of my biggest learnings during the hiring process.

TIM: And what about AI then? So this is the buzziest of buzzwords probably in the history of the world, and it seems as though it's impacting every industry, every domain, in some way, shape, or form. What about in terms of hiring? Have you started to dabble with AI for hiring? Have you seen candidates use AI in the hiring process at all?

MRUNAL: So we don't currently use AI or automation in our initial hiring process, but as I said, I think it will be helpful for the initial CV screening process at least, given the systems are designed really well. But in terms of any skills, I just feel since Gen AI is the major focus on the AI high curve, if I had to hire someone now and needed AI as a skill, I wouldn't specifically look for it because, as we just said, it's a buzzword. AI is not something new, so if a candidate possesses solid, like, fundamental or foundational skills in machine learning or natural language processing, let's say, because we use LLMs, that would be enough for me. So, to give you a quick example of what's happening or what happened in our team recently, when we worked on a CTV match, which is, by the way, our new product that we recently launched, and we developed it using Gen AI, even though the team lacked proficiency in it, once they identified that LLM is something that solves the problem, they quickly learned those new concepts to bring the product to life. So that willingness to grow alongside technology, I think, is the kind of mindset that is needed in today's data roles.

TIM: Yeah, I'm hearing that a lot at the moment, and I think it's one of those points in history where, because the technology is changing so quickly, like really at a breathtaking pace, OpenAI has got their 12 days of Christmas thing going on at the moment where it seems like every day is, Wow, that's quite substantial. and they're just one of the businesses, so I feel like in this environment maybe it's even more important than normal to have that growth mindset, that adaptability, that willingness to learn, because your skillset can almost be it feels like your skillset could almost be redundant in a few weeks; that's how quickly things are changing. and so you really have to find those candidates who are keen to learn how to have a kind of perfect interview question or a particular technique where you try to coax out that, you know, that ability or willingness to learn because I feel like it's It's very easy for candidates to talk about I like to learn. I like to do this and that, but How do you really know you got any secret sauce?

MRUNAL: Oh, I haven't found it yet. I'll let you know as soon as I do, but I feel just asking them how they approach it, so for example, they're working on a project, but they don't know anything about it, like what to use or what not to use, and how do they approach their research process and the different steps that they go through? I think that is a good question. The first step is just to know if that person has that keen desire to grow or ability to learn, but it's very difficult to measure, to be quite honest.

TIM: So we identified this a couple of years ago as our single most important value that we were looking for in candidates: a willingness to learn new things, because we figured that it's a startup, stuff changes all the time, the tech stack changes, our problems change, and our solutions change, so we just needed people who were adaptable, like the current environment, maybe not as extreme. and so we were also thinking about this, what we landed on, which has its pros and cons, was we thought, like, a lot of interviews are very much focused on candidates talking about things they've done and, like, having examples that behavioral style interviews, but I feel like the challenge with those is that talk is very cheap; you can't really validate that they did those things. and it's just a really good storyteller could come up with some very convincing examples of how they've demonstrated X, Y, or Z, and so what we wanted to do was have a way for them to actually demonstrate it, like to do the thing that we're asking them to do, so in terms of learning new skills in our projects that we gave them, these are software engineers. We asked them to build this little algorithm, but in R, like R, the statistical programming language, because we knew none of them would know that because they're engineers that don't write R, and so it was an interesting experiment to see. Oh, would someone say I'm not interested in that? Why would I want to learn AH as a dumb language? That would all be red flags for their lack of a growth mindset, and then if they just couldn't figure it out, like they couldn't download and install the packages, they couldn't run it, and they just couldn't hack it. out I feel like that would have been a bad indicator for their ability to then learn new other things So that's the closest we came to this, but of course that involves giving them a whole project, and it's unrepeatable; you can't scale it across other roles, unfortunately.

MRUNAL: Yeah, but I think with ChatGPT now, that might have made things easier. Someone can just add in ChatGPT, tell me how to write this in R, and get the output, so yeah, it's making it difficult and difficult.

TIM: Yeah, and I guess the other way to look at that is maybe now a candidate's ability to learn new things is for any candidate now easier, like now that large language models have given you a baseline skill in anything, really anything that involves writing, then maybe, yeah, maybe any person's ability to learn something should now be easier than ever if you've got this magical machine to turn your attention to.

MRUNAL: True, but I don't think it's still fully reliable. I know it has come a long way, like since the first chatty PD version and now, but given that it cannot still tell if there's an M in Canada or not, then I don't think it's still at that level but definitely can help a person point in a certain way in a certain direction.

TIM: Yeah, it still feels like at the moment right now things are changing quickly, and that probably the best use case is to use it for something you're already a domain expert in, but it just allows you to do it 10 times faster rather than using it on something you have no idea about at all.

MRUNAL: couldn't agree more, yes, because we need to spot the errors as well as whatever that chat GPT generates, so yeah, having at least a certain bit of expertise to find that, okay, I don't think this was right, and then questioning that prompt, because sometimes when you question it, it says, Oh, sorry, I overlooked it, or something like that. So yeah, definitely.

TIM: Yeah, it needs a prod in the right direction. What about on the candidate side of things? What are your views on candidates using things like ChatGPT or Claude during the hiring process? Are there any steps that you would be unhappy with them using it, or do you think it's fair game now that they could use it anywhere?

MRUNAL: Oh, I'm not sure anywhere, and also it depends on how they're using it. If they're just putting the question or whatever is the question and you're hiring a questionnaire, let's say, and the chat GPT is generating it for you, that's not the right way, but if you're just giving them, like, Okay, this is what I'm thinking, Can you just check the grammar for me? Let's say I think that is enough, so I think there should be the level or the balance between how you are using and how much you are using and for what you're using it because we so even when the person, for example, let's say it used TATTPD and got hired right. Once you get hired, the employer is going to know what kind of person you are compared to what you answered in your questionnaire, so yeah, there should be a high level of honesty in what you do, especially in the hiring process.

TIM: Yeah, it's a tricky one, isn't it? Because some people would say that if the hiring process is such that a candidate can just put stuff into ChatGPT and get an answer out, then either the hiring process isn't that close enough to the actual job, or it is close to the job, in which case why do you need the person? You could just have AI do their job if the hiring process is close to their job, so it's a tricky one, and so would you view candidates as using it in a very brute force way? Let's say, for example, you had an interview with a candidate, and they had ChatGPT open at the same time. and so they were like it was listening to it was doing voice to text, and so it was also giving some answer options for the candidate. The candidate was looking at this other screen and translating the chat GPT stuff back into normal English and answering the questions. If they used it that way, how would you feel about that?

MRUNAL: I won't like it because they are not their own thoughts. Yeah, that's my honest opinion. It should just be who you are. You don't have to be fluent in what you are saying. It's okay to have the ums and uhs, right? But just say what you want to say. You don't have to be grammatically right. but just say answer the question the way you feel it rather than how the machine feels it

TIM: Yeah, I tend to agree with you. I was hiring some people recently, salespeople, and as part of the hiring process there was an application and an assessment. One of the questions was, and this is like a written answer we were asking for, one of the questions was, Imagine it was your first day at Uber. What are the three things you'd need from us to give you the best chance of being successful? So it's a very personalized question where I wanted to know that exact human being's opinion.

MRUNAL: Yeah.

TIM: It was amazing how many candidates still use ship for that, which I found quite surprising 'cause it was a very opinion-based thing. It was almost like sending out a survey. I don't give a shit what church GPT thinks about this. I care what you think you know what I mean

MRUNAL: It's not right, especially if you're a salesperson, because that's your key skill: you talk and communicate, so, yeah, the sales piece of people definitely shouldn't use it, and nowadays I think it becomes a bit easier as well to identify if someone is using ChatGPT to answer because it is written in a certain way.

TIM: Any good use cases of it? For example, one thing candidates are clearly using it for at the moment is crafting their CVs to each job description. Others are using it to apply en masse to many roles that they could otherwise do manually themselves. Are they good use cases, or do you feel like that's still going down the wrong path? What are your thoughts?

MRUNAL: Yeah, no, I think it's still on the wrong part. Maybe crafting for a CV is like different job descriptions. I think it's, again, a thin line how you use it, but for mass applying a lot of job applications, I don't think it's fair to the other candidates, right? who are actually giving time to apply to each and every job, yeah

TIM: I guess the devil's advocate to that would be the tools out there. If you're going to be more efficient with it, why would you want to discourage that? Some people would say, especially if they are in an organization where they were very actively promoting the use of AI to be more efficient,

MRUNAL: I'm not saying you shouldn't use AI; I think I'm just saying where to use it and where not to use it because hiring is very personal, and you're hiring the person and not the machine. As I'm saying, yeah, it should be at least honest when you're applying to jobs.

TIM: Yeah, I feel like we should just have an honest and open conversation about it. If anything, I feel like candidates are probably going to use it, so we should say, Okay, cool, I know you're using it. How are you using it? and then we can start to get to the next level of Oh, using it this way Did you think of writing the prompt this way, or why are you approaching it in this manner? We can start to get past the first layer if

MRUNAL: Yeah, if everyone is doing that, I think that something will have to bring it into practice.

TIM: One outcome of candidates using AI seems to have been that—and this is based on anecdotal feedback, but a lot of it over the past two months—is that the CVs tend to start to look like each other because AI has written them based on the same job description and optimized the candidate's CV, and then the CV is looking more similar to each other. and then each job has way more applicants than normal because I think probably two factors are one is candidates are applying en masse with these tools, and then I guess the suppressed market conditions where it's just that there are more people looking for jobs in general, so there's this weird thing where now a candidate is probably looking at LinkedIn going, Oh wow, there are 500 applicants already in a day. Are you kidding me? I thought I had to apply for 50 jobs to get one interview. I have to apply to 500 to get an interview, which is then causing probably more people to apply in this weird vicious cycle. How do we break out of this? How is this going to end? This seems like it's pretty flawed, no?

MRUNAL: Yeah, I don't see an end to this if 10AI is evolving the way it is. I think it's going to become worse if we don't try and put something in it, maybe have something like plagiarism in it like we have for our master's or when we submit our essays and there's plagiarism. Okay, maybe we can introduce something like that to see, okay. The skills will be similar to other CVs, but maybe the way it's written, as you said, CVs are exactly similar, so maybe that could be a better way of identifying if you're just copy-pasting from somewhere, and this is not exactly who you are, so plagiarism tools, I think, could be something that can be introduced in the hiring process.

TIM: Yeah, perhaps, and I guess what's going to happen for companies is they will have to all automate the screening step because nobody can read a thousand applications. That's literally the number that some of these jobs in America, in particular, are getting: like thousands in the first day. So I assume they're going to automate this mainly with some kind of AI CV screening tool. Maybe they might do a skills test as a first screen, but probably they'll screen the CV first, which then means it's going to be like Chachapiti against Chachapiti in creating the CV and then screening it, and yeah, who knows where this will end because I feel like personally if you are using Chattopadhyay to write or augment your CV, surely it's going to be less truthful to your point. like you're going to need some kind of check and balance to make sure people aren't just making up complete nonsense and sending that off

MRUNAL: Exactly, yeah, it's going to be difficult, but I hope it doesn't introduce a lot more stages to the interview process because right now, for large companies, it's seven rounds and like that. I hope it doesn't go to 15 rounds because it's not fair to both the employer and the job seekers. So yeah, let's see how it goes. I'm interested and curious to see.

TIM: Yeah, so I wonder if in the short run, companies are going to be inundated with all these applicants they won't be able to do it manually. So, like, I spoke to one company last week that said they put up a job ad, and they got 1100 applications. They read the first 150, and they got five candidates they liked. They're now interviewing them. That's it. The next 900 they didn't read because they couldn't, and they already had enough good candidates, so I think there's going to be, unfortunately, lots of people missing out and maybe not getting a fair chance because there isn't the technology in place yet to deal with the volume.

MRUNAL: Yeah, no, you're right, and it might take some time. Yeah, it should happen soon.

TIM: Yeah, and I guess that makes sense that the candidates themselves might be early adopters of AI. It's going to take any company, even a startup, longer to make decisions to implement a whole new technology or way of hiring, whereas a candidate can just start picking up any tool they like anytime they want.

MRUNAL: Yeah.

TIM: I'm changing the topic a little bit, so candidates are in this market at the moment, which is that they're facing a lot of competition. They're looking at these job boards like LinkedIn, and they're seeing, Oh wow, a thousand applications? Wow, that is a lot of competition. Now a lot of them are still then probably reacting to that, saying, I'm going to have to apply to more roles. So they're going to say I was going to apply to 10; now I'm going to apply to a hundred. I'm going to automate my application process to deal with this massive amount of competition. That's one way to do it, and that's to go through the job sites as a channel, but there's another way to do it, maybe, which is to directly try to get a foot in the door through some other means, maybe through your network or reaching out to the hiring manager. What are your thoughts on candidates doing that? Would you be happy for them to do that to try to get a foot in the door directly with the hiring manager through it, let's say, a nice LinkedIn or email message, or do you feel like that's just them trying to dodge the queue that they should be a part of?

MRUNAL: Yeah, I think they should follow the application process because it's there for a reason. Hiring managers have their own workflows or the way they want to evaluate the candidates, and if you're approaching it through a network or other way, it might just introduce confusion. You might have a chance at the application process, but because you are approaching it in the other way or going around, you might not, so yeah, I think they should follow the application process, but having said that, following up in a week or so after the application Or whether you have interviewed, I think that's fine. Because it just shows the eagerness that you want to work for the company, and it's not something like one of the applications that you apply from the bulk, so following up, I think, is a good strategy, but again, it depends on how you're following up, because one time I received a message on LinkedIn with just one line saying, Do you have any job openings? That is it. No context, nothing like, okay. This is my name; this is what I'm looking for; this is my CV. Nothing. So asking politely and following up, I think that's a good way of showing interest, but yeah, just be polite. I think that's, I would say, because it shows your soft skills as well.

TIM: Yeah, I'm sure any hiring manager listening out there would have received some pretty low-quality LinkedIn outreach in the past few years similar to what you've described there. Yeah, we would get off with us just, Can I have a job?" with a random CV from some random person I've never met? from some country where we don't hire people just like completely irrelevant to us and our needs as opposed to Oh hey, I noticed you're hiring for a very specific role: software engineer. I noticed it required X, Y, and Z. They're all skills I have. I just finished my last role. I live in the same city as you. Here's what I could bring to the table in some kind of personalized way: would you be receptive to that kind of approach if it was like

MRUNAL: Yeah, for sure, yeah, yeah, that's good, but again, don't use ChatGPT to do that. Yeah, but no, definitely, if you're introducing yourself, point out the skills that are similar in the GED and that you have and how you have used them, then definitely

TIM: And what about for yourself? Imagine, God forbid, you were in the job market in a month, and suddenly you were facing this stiff competition. Would you still apply yourself through job boards, or would you try to perhaps leverage your network or find some other way to get a foot in the door?

MRUNAL: Job spots definitely—I don't want to be saying one as a hiring manager and the other thing when I'm a job seeker. No, I would still go the normal way, and then, as I said, follow up after a week or so just to check in on where my application is at.

TIM: You would, what's the phrase, you would practice what you preach, basically.

MRUNAL: Exactly, yeah.

TIM: What about thinking bigger picture now about the hiring process for data roles? Is there anything in particular you would like to redesign? If you could click your fingers and just radically change it, what would you do? What bit would you attack first? What would be the biggest improvement to be made, do you think?

MRUNAL: Attack first. I think the first step will always be the initial CV screenings. I don't think that's going to change because, as we said, the number of applicants is increasing day by day on just one job post. And the best way to see, okay, if there's one applicant who takes most of the boxes, then let's move on to the next one. So I think the CV will always be the first thing in the hiring process, but I'd love to hear your thoughts if you think otherwise.

TIM: I think the CV personally is not that useful because it's just someone's opinion about themselves; at the end of the day, it's not really validated. I can say that I'm a rocket scientist on a CV and go and apply for a job with SpaceX; no one can stop me. I reckon I could probably get an interview at SpaceX just by writing a really great CV that's made by me.

MRUNAL: Interesting, yeah, yeah, I wasn't thinking that people were lying on the CV perspective, but yeah, I think, yeah, please, guys, don't lie on your CVs, yeah, yeah.

TIM: No, it's a very short-run strategy that, yeah, would leave me very embarrassed in my first interview at SpaceX when they're asking me stuff about physics and I don't know the slightest thing about that, so I'm not sure I'd get through the hiring process, but I think I could get a foot in the door, but yeah, I feel like the problem with a CV is that so many problems So one is, yeah, someone's opinion about themselves; it's not accurate, and I'm sure we've all had experiences of interviewing someone who looks amazing on paper only to be very disappointed when they don't seem to know anything that they claim to know, but then there's the bias step, which we discussed earlier, around people's names and ethnicity and gender and religion—all these things that come across on a CV that, at least if a human's doing the reviewing, you can't really get rid of easily. And so I feel like there's an opportunity for a more objective screening step. The problem, though, is I think data quality, because we only have a CV. There's so much more we want to know about a candidate, but we just don't have that data yet at that point, like we would love to know their personality, their intelligence, their skills, and how well they fit the role, but those are only data points that are collected later at the moment. So I feel like we need that data, but we need it earlier.

MRUNAL: Yeah, now that I think of it, after discussing it, I feel that what can be changed is maybe rather than having no specific formats for CVs, we could have questions or a universal CV format with specific questions, and those things should be in there, and that way we could collect those data points that you were talking about. and I think that could be a better way than having a system automated based on those data points because now every CV in the world is going to contain those data points rather than having people design CVs the way they like. They don't usually contain all the information that you need, so maybe having a structured format might be helpful.

TIM: Yeah, and I think in Europe there is Europass, is that right? I think it's a CV format that is becoming quite common now. I'm not sure how closely they dictate the exact form, but I've seen a lot of them over the years. I'm sure it's Europass, or is that a train ticket? Maybe I'm getting the two mixed up.

MRUNAL: Yeah, yeah. a train ticket or maybe

TIM: Yeah, it's a European Union-style CV, Europass. There's some connection to the actual EU in it. I think they have some standardized template that people use, but yeah, it's not like everyone uses it, so it's still not a hundred percent universal. I feel like still the issue, even if we had that, is still the validation bit is missing because maybe the validation and the subjectivity because I could say I've advanced SQL skills. What does that mean? I've advanced Excel. Is that a pivot table or is that VBA? It's down to your interpretation a little.

MRUNAL: Yeah, then maybe at some point those skill test skill assessments on LinkedIn might be helpful to check if the skill that the person has mentioned is actually the skill that that person possesses, so yeah, maybe something like that.

TIM: Yeah, I think there has to be some kind of validation. What would be an interesting business, actually, is a company that sits in between the candidates and the hiring companies because pretty much every company wants to know a very similar set of things about candidates, but it's very annoying for candidates to have to go and repeat yourself 50 times to 50 different companies explaining your SQL skills or demonstrating a certain value or a soft skill. So if there could be some intermediary that validated all that data and then shared it with companies,

MRUNAL: Yeah.

TIM: That would save a lot of time, but I'm not sure I'm thinking a little bit, yeah. We don't do that bit maybe one day, but yeah, we don't sit there at the moment, but yeah, hopefully someone solves that problem and finds a way to make money from it because it would save everyone a lot of time.

MRUNAL: Definitely, time is money.

TIM: Exactly what about this curly question If you could ask our next guest one question about anything you like in the world, what would that question be?

MRUNAL: Now this is something very basic, and we discussed in our introductory call that SQL is an essential skill to have if you're hiring for any data role, like data analyst, data scientist, or data engineer, because, as I told you, I have heard mixed opinions. My opinion is that it is extremely important because whichever role you are in, you have to do some exploratory analysis. In SQL, right, just to get to know the data, so I feel it's extremely essential, but what I noticed was when we did our hiring the last time, only one out of 10 candidates had SQL on them, and they all were recent graduates, but graduating from a data course, I don't know what has changed between the time I graduated and this recent one, like what the course is changing. Maybe the academic courses are prioritizing more advanced skills, but I feel the basics, the foundational fundamentals, should never leave, because they're still important in practical life, so that would be my next question, or my question to the next guest.

TIM: We'll level that at the next guest, and yeah, it's an interesting one and so fascinating that the qualifications are these university graduates from like undergraduate degrees in data science. They have not been taught SQL cool.

MRUNAL: Yeah, undergraduates or even postgraduates, and they knew so one of the candidates knew no sequel but no SQL, so SQL is an essential skill for us, whatever work we do in our day-to-day life, so my first question, or one of my questions if we are talking about SQL, is do you know the difference between a union and a union all, which is very basic? and they're like, Sorry, we don't know, and I was like, Okay, it's yeah, and given that this is something mentioned in the job description that SQL is essential, I feel it's courtesy from the job or the candidate side as well that they should read something about or mention that, okay, I'm taking a course to build up on it. and that again shows that ability to learn or the growth thing, so yeah, it's very interesting and shocking for me.

TIM: Yeah, I would also be shocked with that, and yeah, I personally feel right now that it is absolutely essential for any data role. I wonder, though, if that is changing quickly, where for all coding languages, not just scripting languages like SQL, a large language model could now, maybe not now but very soon, be doing all the coding for us, and we're just going to interact with it using natural language. Then maybe at some point it's going to be redundant, but that point isn't yet.

MRUNAL: Yeah, I don't think, at least in your future, it's going to become redundant.

TIM: No, it's the language that doesn't die. I think it was IBM or one of those companies that created SQL in the sixties; like, it's really old.

MRUNAL: very yes, exactly, and if it hasn't died yet, then yeah

TIM: Yeah, it's like the spreadsheet; it's just it keeps living on the cockroach of the business world.

MRUNAL: Very well put, yeah.

TIM: Murnell It's been a great conversation today. I've really enjoyed it. We've covered a lot of ground, and I'm sure our audiences enjoyed your insights and thoughts.

MRUNAL: me too, thank you so much again. Thank you for the opportunity and reaching out to me