Episode 13

Praveen Tomar on the Future of AI in Data Science and Ethical Hiring Practices

TIM: Praveen I'd love to get your thoughts on this. AI is obviously having a profound impact on the world already. Lots of things are changing; probably everyone in every job, especially knowledge workers, are thinking about what's going to happen to my role? What will my role look like in a few years? If you had your crystal ball, what do you imagine data scientists, data engineers, and data analysts will be doing in five years time?

PRAVEEN: That's a really good question, and I would say one of the pressing questions nowadays, because ultimately everybody wants to do something with artificial intelligence or with generative AI, and data is the first step or the stair to reach the maturity of AI, right? And we all know that, right? We love Excel sheets; we love SharePoints. Most of the data resides on those in not very good shape and size, but over the top we want to build some kind of AI. Why is there so much need for bringing some kind of AI into the organization? Because everybody believes, and since ChatGPT brought the—what do you call it?—the power of AI to the common people, right? So now everybody wants to improve productivities in some way, in some form or shape, right? So the data team is the first go-to per team, or the people, right, who can do something with AI. There's nothing called you can't hire an AI engineer. All the AI engineers, their title might be AI engineer, but their background must be data specialist, data automation, data engineer, data scientist, or something like that. and without having nice and clean data, you cannot have AI, so in the next five years, what I see is that most of the data specialists, whether they could be data analysts, data engineers, or data scientists, are upscaling their understanding about the various types of AI tools, platforms, and models. and challenges and risks associated with AI so they can communicate and articulate back to the business, Hey, you want generative AI like in my team. Many business people come to me and ask me. Hey Praveen, can I have Gen AI on my team so we can say, Yes, of course you can have it, but do you have clean data? Do you have structured data? Do you have data that can be fed into training the AI models or the machine learning models? So understanding about AI and machine learning and what the risks associated with it are and how that can percolate back and go back to the data or the data quality that they have and articulating them and identifying the risk and explaining them to the business, which I'm seeing it and learning new tools and technology As a data specialist, there is no escape in terms of learning new tools and new technology like Azure Synapse or new tools from AWS or new data platforms or Microsoft Fabric from Microsoft, right? So for those who are Power BI engineers, now they have to upgrade themselves. So they do not have a choice; they have to upgrade, but at the same time they need to learn AI and machine learning skills. This is the fundamental thing that I'm seeing in the next five years.

TIM: If you think about the day-to-day tasks of either an analyst, an engineer, or a scientist, are there things that you view as like low-hanging fruit for large language models or other types of AI to help with, for example, the preprocessing of data, data cleansing, aggregation, or perhaps something else?

PRAVEEN: For any organization, I feel that data quality and how data is coming into the data pipelines, data acquisition, and how good that data acquisition platform we have are key challenges that I am seeing. If you do not have a nice structured way of getting data in the organization, then you have a lot of data quality issues, and you need to do a lot of human intervention required to fix those issues. and there are probably very few little automation solutions available that can identify data quality issues and fix them automatically. You need data engineers to identify those issues right and then fix them, so lots of manual intervention and human in the loop are required, but if somehow we can get a nice way of getting data into the organization in a structured or at least semi-structured way So you can do some nice things only on the structured data; if the data is unstructured, it's a nice long, lengthy paragraph of free text. It's as good as email; you just need to go and read and do the comprehension of those texts, so these are the key things that I'm seeing day to day. and also upgrading and updating a skill set is, I would say, nowadays, probably the best practice if we cannot assume that everybody knows everything, so attending conferences and going out and getting the knowledge from outside how other teams and other organizations are dealing with data quality issues and what kind of data governance structure they have

TIM: Thinking now about hiring, I was always struck when I looked at the way recruitment was done that it's often quite gut-feel-based, quite intuition-based. Not a lot of things normally get measured as part of the process. I'd argue it's generally not done in a data-driven way, and it's certainly not traditionally done in an advanced data science way. But what about your experience so far? Have you seen any, perhaps like small-scale efforts to make hiring a bit more data-driven and a bit more objective?

PRAVEEN: Yes, so it's a very good question and one of the pressing challenges in terms of the data leaders is how to hire the right resource, and the hiring process in itself is quite generally long-winded, right? Even if you get the right resource to get them on board, it takes a minimum of three to four months of time. and in my last organization they used to have a very good practice called pre-screening questions so the candidate can record their answers against those questions, and those are based on just two or three questions, not too much, so these are the open-ended questions when we ask the candidate. to share their experience about this real-life problem, right? And they don't have to do anything, do any coding; they just need to articulate how they will approach this problem, so rather than focusing on theoretical technical knowledge We also wanted to see how they are applying their knowledge to real-world scenarios. quick video Record your video with five minutes or three minutes of answers on these two scenarios and submit it. and if hiring managers see yes, the answer is good, the approach may not be the right answer. This is how a data analyst, data engineer, or any normal person with good reasoning will approach it in this way, so that gives them the first step to shortlist the candidate, even making an effort to go into the detailed resume right. So that's a really good step, which I believe will sort out a lot of noise from the first step itself, right? Those who are serious candidates, right, they will make an effort to give a very good answer and record that video and share it in response to those top two or three screening questions, right? and otherwise

TIM: And

PRAVEEN: We are all human; we just wanted to do shortcuts right and copy, paste, and submit it and, yeah, to avoid those things.

TIM: Yeah, a video brings it to life. I'm wondering if part of the approach was to standardize what you thought a good answer was and have a mechanism to grade the responses from the candidates.

PRAVEEN: So there are various factors, right? First is with the responses of those video recorded by the candidate, giving you two or three things as a hiring manager: one is the communication skill, confidence, and problem-solving approach, and also a little glimpse of their technical expertise, right? So you may not be able to explore full technical expertise, but that will give you a lot of confidence, and still I would suggest writing the technical skills that can be upgraded, which can be learned, which can be fine-tuned, but something which is problem-solving communication leadership confidence. This can also be built, but it takes a lot of time, right? It takes efforts, years of efforts, right? So, you know that with the hiring process in itself, right, okay, where some guy or some person is, it is very good technically, but he's not able to articulate the problem in a right way or the solution in a right way. and in our real life we have to interact with the non-technical stakeholders, business owners, or the process, and we need to articulate a solution in non-technical language, right, so they can also understand, like the head of the business; they won't understand when you say Hey, okay, I've run some Python script, and this is the output we have got. What does that mean? You have done some hypothesis testing. What does it mean? It's not the data that is showing there is no abnormality or no abnormal data is there, and it means the data that is showing in terms of trend upward trend or downward trend is the true representation of your business processes or the business outcome. So it's ability to explain the complex technical issues in a language that everybody can understand; a stake business can understand it in simple language, right? So that's that; that glimpse we can get from those videos.

TIM: Yeah, I think a video really brings it to life in a way that pure text wouldn't because you just get to see that there's

PRAVEEN: And Stanford University, they started this in their MBA essays, right? Remember a few years back we used to write 1000-word essays for our MBA applications, so they said, Okay, people, we do copy from here and there and Google; we go to the experts who help us in terms of writing personal statements, right? and they write very well; they are professionals, but when you record your video, you are yourself; you're a genuine person, right? You cannot borrow someone else's knowledge in your personal statement, so they started this one minute, and also the video gives you the information about the candidates energy level and enthusiasm, right? which cannot be marked, which cannot be There is no parameter in tech interviews to assess this, that is, how enthusiastic the candidate is about this new role and how likely it is that, if the role is being offered, they will be able to join it; otherwise, some people are very good, but they are just exploring to see whether they can get a better offer or not. So a lot of other inputs you might get it, yeah, so that's what I was thinking.

TIM: Most hiring processes would start with a candidate taking their CV, uploading it, and applying for the role. As you mentioned, adding in a small secondary step of that short video screen gives a lot more information than a CV alone, but for the companies that are relying on CVs, do you feel like that's not a great predictor of who the best candidate is? Do you feel like companies over-rely on that instrument in the screening stages? Perhaps there's another data set they could use to decide who to interview.

PRAVEEN: No, that's a really complex challenge, and when we advertise a job vacancy or the role, generally it's 500 words or maybe 1000 words or 1200 words in terms of job description, and of course it does not cover everything, and there are so many other things that you have to do on a day-to-day basis, right? which you cannot; you cannot write everything, and otherwise the job JD will be 10 pages, and nobody will apply. People get scared, so also, yeah, cool, I would encourage we need to mention somewhere in the JD itself, Hey, even if your skill set does not match with all of those items that are mentioned in the JD, if it is 60-70%, still go ahead and apply; try to apply. and it is also an opportunity for the candidate to have an informal chat with the hiring manager that's not an interview just to understand the overall scope of work. What do you call it? The recruitment seminars we need to conduct, which when I was working at Deloitte, they used to do it right, so they give the glimpses of the team, the hiring team, what technology stake they are using it and what And some insight peek peeking inside their team that, okay, probably the candidate is going to join or going to join this team, so that helps a lot, and yeah, and you know, giving them an opportunity to have an informal chat rather than sending a CV and then seeing, "Oh, there might be a very good candidate. But we rejected him or her because he or she did not mention something specific in the resume, right?

TIM: And what about once we get to the interview stage? There are different types of interviews; most commonly, you might have a cultural fit interview, a behavioral interview to check softer skills and experience, and a technical interview to test those technical skills. I tend to hear, I'd say the majority of companies say that they would favor the softer skills. because they feel like the technical skills are a little bit easier to teach, a little bit easier to pick up, and they really drill down on the cultural fit and the soft skills, but you could hire the friendliest, nicest person in the world who's a great communicator, but they might not know anything about data at all. Do you feel like this mentality is sometimes actually hurting the ability for data teams to deliver high-quality work because they might have a lot of good talkers and friendly people but not necessarily those who can actually do the work?

PRAVEEN: Yeah, Tim, this is really a good question, and my answer would be the ideal of how we can strike the right balance of both. We cannot have a very nice, friendly person who is excellent in communication but not understanding anything about data, or we cannot have a person who is excellent, superb, like a data one world's top 2 percent data scientist but not able to articulate properly. So we need to strike a balance. You can have 80% technical skill set and 20% okay, still can be improved, and similarly in communication, if a person can articulate basic information, that's fine. We don't need a PR agency or a news journalist kind of person, right? So striking a right balance and mindfulness in terms of nobody knowing everything, right? So that would be the ideal candidate for me, and this is what firms should be focusing on, right? So not going on extreme left or extreme right is technically extremely good, but not being able to articulate the basic problem or basic issues or the solution in a normal language We cannot; it's not very useful in either way as well.

TIM: It's about having that balance but also not expecting a unicorn.

PRAVEEN: Yeah, absolutely. You are asking for Superman, and then you have to pay for millions of pounds if you want Superman to do very good, excellent, superb communication on a scale and excellent data scientists. No, we don't get those kinds of people. But the important thing is that the ability of the candidate or willingness to learn something, which is lacking, right? That is very important, which is generally missed out, so even in an interview when the hiring manager or the firms realize, Okay, the candidate is very good, but it's lacking in certain skills. maybe communication or maybe some technical skill set, but what is the willingness for the candidate to also show what you call the portfolio assessment or the CV assessment? how regularly a candidate is learning some new course or engaging in learning something new that gives you a lot of good indication Yes, the person is in, and over the last two or three years, has learned various other things, so the person is having a genuine intent to learn. If something is lacking, it can be trained and can be upgraded. Okay, so that's a very good insight that can be drawn from the candidate's resume and can be asked as well. but in an interview, if you ask general candidates, I agree with everything, but if you ask them politely, Hey, what are the last one or two new skill sets or something you have learned? and not mark them against them, somebody might say, I've done 10 new courses in Azure and five new courses in AWS. Okay, but the ability to have a continuous track record for the last six months or one year, how many new things they've learned, and also noting down those keywords in the responses that when the candidate says, Hey, I came across a new platform, a new solution, and I have learned, and I've done these certifications, right It's very good evidence of candidates ability to learn, their ability to be trained, and their ability to upskill themselves.

TIM: One thing I'm really interested in is making hiring fairer. Ultimately, this podcast is all about objective hiring, and I feel like a lot of the traditional methods of hiring are very potentially biased, so just to unpack that a little bit, a CV contains a lot of information that really companies don't need to make a hiring decision. So straight away the name reveals someone's gender and ethnicity; you could easily get their religion based on their school; you could get their socioeconomic status based on where they live; some markets people put a photo on, so suddenly you know what they look like in the Philippines, and it's very common just to put straight at the top: Roman Catholic married. So you've got the marital status, you've got the religion, and there are so many data points on there that are just completely irrelevant to deciding who the best candidate is, yet until, let's say, very recently, most hiring processes would start with a human reading that CV and trying to make a decision despite being bombarded with all that noise. And so I feel like a lot of hiring is set up like that, but it's just almost inherently potentially biased. Have you seen any interesting ways that companies are trying to make hiring more objective and fairer?

PRAVEEN: Absolutely, and a very good question again, and we have a solution for this in UK civil services. The hiring process is completely anonymized in terms of candidates names, dates of birth, or anything that is personally identifiable. It's completely removed, and if a person deliberately tries to enter that kind of information, their application gets rejected. Hiring managers only get the employee candidate ID and the skill set, and we also don't encourage you to write your company's name like somebody has worked for Microsoft, Facebook, or Google; the brand in itself is having a power to influence decision-makers. university degrees Like somebody graduated from Oxford, right? As a hiring manager, we all know that okay, people coming from Oxford and Harvard, they are the best of the best, right? So you get—you tend to get biased, right? And so avoiding those things, we do anonymized recruitment, name-blind recruitment, or no personal information is being captured and encourage candidates to not give any personal information. And also, along with the CV, we also ask for the personal statement and motivation of the candidate, why he or she or they are applying for this position. This is very important. I personally, if I have to do a screening basis on the CV, I would say I would give 60 percent to the personal statement and motivation why the person is applying for this job and 40 percent to their profile or the CV, right? So the personal statement reveals so many good things about the candidate, such as his intention and how he can summarize his entire work, or the key work that he or she has delivered in the past, let's say, in the last year, into those examples in the personal statement, which are matching with the key requirements of the JD, right? So it must have done some good amount of homework before writing the personal statement. If you just simply copy-paste from your previous job or from a previous application, it won't solve this. This also reflected how the candidate how much a candidate is serious about the job. If you are serious, you would really want to burn your midnight fuel and spend some time, quality time, one hour, two hours, to write a quality personal statement. and I personally feel that I get influenced a lot by a good quality personal statement rather than a very good quality CV.

TIM: And have you noticed, so these statements, these are the ones that are written, or these are the ones that are like a video recording?

PRAVEEN: Both video recordings are the questions that are related to the real-life scenarios problem, right? Real-world scenarios where the candidate is explaining their approach in a simple two- or three-minute video. Let's say the question is that, for example, in an analytics dashboard you have a lot of what do you call it, spikes in the data. So how would you approach starting out this issue? How would you approach solving this issue? We just wanted to hear the approach right. not the technical skills in Python or R or C# or anything else, right? So that gives the problem-solving approach, but the personalized statement gives the motivation that is why somebody is motivated enough to apply for this role.

TIM: Have you noticed a trend in the last year or two of candidates feeling like candidates might have used ChatGPT as part of their preparation? If so, what are your thoughts on that? Do you think they're going to use that at work? Of course they should use it in the hiring process. Or do you feel, Ah, this is about motivation. I want to know what the candidate thinks. I don't want to know what ChatGPT thinks. What's your view on ChatGPT, specifically for those motivational kinds of questions?

PRAVEEN: I would suggest the candidate not use ChatGPT. Even if you make grammatical mistakes, it's fine; we are human; nobody's perfect; we do mistakes. spelling mistakes, grammar mistakes, do it reasonable amount of mistakes are acceptable; it's fine. Nobody expects you to have a perfect the perfect resume, the perfect personal statement, which also reflects you are a human, right So it's fine, and the motivation should come from the role that you are applying for, right? Why this role is motivating you, right? To apply for it and use ChatGPT for fixing grammar or spelling issues. I'm fine with it, right? I don't mind it, right? But it should reflect your real motivation, not ChatGPT's motivation. because he or she ChatGPT is not going to come and work for you, right? You have to work, and you should not get frustrated. Let's say even if everything goes well, happy path scenario, you get the offer, you join the organization and the team, but you are not feeling satisfied, so you are asking trouble for yourself and for the team as well. So my suggestion is that you use it cautiously and judiciously, which is absolutely needed, and then only use it.

TIM: In my experience in seeing candidates answers to thousands of different problems over the years, it's sometimes quite obvious when they've used ChatGPT, at least in versions up until now. Maybe it'll become indistinguishable soon, but at the moment it's still often quite obvious it uses this kind of over-the-top, flowery language. It's very verbose. It's got a certain tone to it that you can just pick up naturally, but I'm interested to see if that tone carries forward to the videos. Do you feel like you can intuitively sense if they're recording a video that they have read it from ChatGPT, that it's not really their true words?

PRAVEEN: So in the recorded video, it's not the candidate who has to do the timing of the video, so we provide the link as soon as you hit the recording button. In two minutes time, the recording will end, so you do not have time to look anywhere here and there in your notes or chat. GPT You have to be your authentic self so you can prepare, and there's no harm and nothing wrong in preparing well for the answers, and yeah, so if it's unlikely detectable, right, in the videos, if somebody is prepared to answer using ChatGPT, Yes, if they use it fine, but is that same thing reflecting their subsequent round of interviews, their personal statement, or their CV? Or when they are across the interview table, is the same thing being reflected right? because the hiring manager won't forget how to answer this problem in the video and how he or she is answering right now, right? So there will be common sense, which needs to be used.

TIM: That's a really interesting point, so you're looking for almost like a consistency in how they respond to certain things throughout the hiring process because, yeah, you're right; if they were disingenuous or lying or used ChatGPT, they'd forget that they'd be almost caught out in their lie. It'd be very hard to maintain; it wouldn't it? So a sign of their genuineness is the consistency.

PRAVEEN: Yeah, so these are some of the good initiatives I would say, and also the informal chat option of civil services provides the option to have an informal chat with the hiring manager to understand the role, team structure, growth prospects, and training and learning opportunities. Those are also very vital pieces of information for the potential candidate, right? It's a window for them, and also if they can join the recruitment webinar or seminar, where we are talking about the role, the team, what kind of problems we solve, what kind of work we do, and what technology stack we have. These are the very good opportunities for candidates to look into the team organization and the role; they will be truly motivated when they apply, and this they can link to their motivation with their firsthand experience of interacting with the team and hiring manager.

TIM: I'm interested in this step, actually, because this would not be a common step that private companies would have, so do you have to qualify to a certain stage to be able to have that chat, or could any applicant somehow get an informal conversation with the hiring

PRAVEEN: No, I think everybody should get the chance. Having a 10-minute informal chat is not a big problem, right? And that is a win situation for both the hiring manager and the candidate as well, and everybody does not dare to go and do the informal chat right, and we need to encourage them because in real life when they In a happy path scenario, when they get selected, they join, and they have to talk every day with their line manager or hiring manager or the team members, so people have to open up right themselves, so if they open up, yes, I want to know more about him; I want to know the location; I want to know about the prospect of learning new tools and technology in your team. What is the training budget you have? What about the certification budget you have? Let's say I want to do advanced machine learning algorithm training from MIT, and that is costing around 3000 pounds. Would your company be able to provide that funding for the certification and training? So that's a win situation for both because the hiring manager will understand, Okay, this person is motivated. He generally wants to upgrade himself or herself and wants to learn. And if they have funding, they can say, Yes, we have funding, so that's a very good match, right? To know each other, I don't want to draw an analogy loosely with a couple who are dating first, right? So they go talk about each other, right? They talk about their aspirations, so that informal chat should be extended to everybody who wants to know more about the team role profile and growth opportunity.

TIM: That's interesting, and I'd like to play devil's advocate a little bit and maybe continue the dating analogy as well, so the last 20 companies I've spoken to in the last few weeks have all complained about being inundated with applications, like hundreds, if not thousands, of applications, and so if their hiring managers said, Yeah, okay, like I'll have a 10-minute chat. Anyone can have that with me. That's all they would do for three years is have 10-minute chats with candidates, or, yeah, for the dating analogy, a girl on Tinder who gets a thousand swipes. I don't think she wants to go on a thousand 10-minute dates with dudes. Is there some dynamic that's different in the civil service that means that a hiring manager's calendar isn't blown out with hundreds of different meetings they're having with all these candidates? Is there some fundamental difference in the way it works?

PRAVEEN: This is a really good question, and it's a very good problem to have for the hiring manager if they get 1000 applications for the advertisement. Honestly, in my last three years of experience, whenever I published any job vacancy, I usually got 50 or 60, and the best of the best canary scenario was 75 or 80 applications, right? and by looking at the quality of applications, 50 percent are just filled hastily right in without spending half an hour or 20 minutes of their lives, right? So you can reject them outright, and like in a personal statement, somebody just wrote one line about how we can shortlist him, right? how we can do justice with other candidates right and it's a good problem to have, and of course one person cannot lead the recruitment process; it is not led by only one person. We need to have a panel of qualified panelists in the interview—not only the hiring manager but also other people in the team who are available to have a 10-minute informal conversation. with the candidate, potential candidate, right, so the workload will be spread, and that is also an opportunity for other team members, especially, let's say, junior team members who want to upgrade themselves so they can have a candid chat with their colleagues as well, right? And also the other managers, those who are your counterparts in other teams, you can also make their bandwidth available right to Hey, I'm hiring a data engineer, but this data engineer works for your data, so if you want to be part of the recruitment process or if you want to be available to have a chat with five candidates for 10 minutes, we need to absolutely strike a balance here. We cannot have a chat with 1000 people, but if generally there are 20 good candidates who reached out or 10 good candidates who reached out for an informal chat, we should extend it. TIM: It's a really interesting natural experiment then. To then see what proportion of candidates are motivated to reach out for that chat, like you would think if all the candidates really wanted the job and they're all desperate for an interview, everyone would book that 10-minute chat. So it's very telling that only a small percentage, or a relatively small percentage, do, and I wonder if those candidates then almost have a bit of an advantage in the hiring process because they've managed to build a little bit of rapport with the hiring manager and maybe extract some extra information they could use in the hiring process. If you noticed, I don't know, you might not have enough data here, but if you've noticed any correlation between the candidates who had that extra chat and their success in the hiring process,

PRAVEEN: This hypothesis, and that's why I, in my previous answer, said we need to have a hiring panel, not just one person, right? So somebody who might have chatted with some other manager, right, in the data team, so let's say head of the data science, not with, but the role is in head of the data engineering team. So he or she might have a reputation built up with other people in the department, but the role is not with the other team, right? So that's why we need to have a pool of the talented people who are very good with interviewing, right? So it will not be too much pressure on the hiring manager, and also the candidates will not have the unfair advantage in terms of they've already built some kind of reputation or first impression with the hiring manager, right? They might have an informal chat with other line managers or other So we can minimize it; can we eliminate it? The answer is no, but can we minimize it and avoid it? Yes, that can be done, and also we need to encourage the candidate, the hiring manager, and the hiring panel to do the recruitment webinar so you can give a very high-level overview. 20-minute webinar or 30 minutes So, can you? It's a window of opportunity for candidates to ask questions and give the overview of the JD team structure or the structure of career prospects and other facilities in the organization, like the cycle club or bicycle club or the nice cafeteria, so these are the things to attract the nicest and brightest talent in your team.

TIM: When I've investigated the literature on what predicts on-the-job performance, it is quite clear that some things predict performance better than others, and so it turns out basically that things like the number of years of experience a candidate has, their age, unstructured interviews like your sort of pub test interview, your gut feel interview, your coffee chat, and those kinds of processes carry very little predictive power on whether or not the candidate will ultimately be successful once they've come into the role. and things like IQ tests some aspects of personality quizzes, job skills tests, and then structured interviews where you measure things in a very specific way tend to predict performance really strongly. Now what I find interesting is that companies tend to do the former and not the latter. So there seems to be a weird disconnect between what has been studied academically and what we know to be true apparently and what companies actually do or organizations do. Do you have any views on why we would keep doing what we keep doing, things that don't work? Basically, any thoughts on that?

PRAVEEN: Yeah, this is interesting. In fact, I think so. Yeah, so not the traditional question-and-answer kind of format interview; they are more predictable if you have data from a hundred interviews or the candidate who came through the proper traditional way of question-and-answer format. You can predict more about their performance and outcomes rather than people who came through informal chat or through some other things. So I feel the later one is Or maybe the combination of both would be the ideal right You cannot have absolute left or absolute right again, so making a decision about the candidate in a formal Q&A format of an interview, plus informal, and then you are judging that, okay, about the candidate, the role requirement is maybe a lead data engineer or lead data scientist. but 50 percent of the time this person has to deal with the stakeholders communication right and the project management so that Q&A format Yes, you can gauge it, but when you do informal chat, you get to know more about the candidate as a person, as a human being, right? Generally, in Q&A, when you are doing a technical proficiency assessment, analytical skill assessment, right attention to detail, right that all those can be captured in the traditional Q&A format, but there is something that cannot be captured, which is maybe before you make an offer, you should ask for Hey, can you have one more informal chat? Right, let's talk about something else: something about football or something about cricket. So I do remember a long time back I had an interview with McKenzie, maybe around seven or eight years back in Washington, DC, so they asked me about it, right? Okay, can you join me? One of the partners in McKinsey invited me for the dinner after having a formal round of interviews. Can you join me? So he knows that the role that I'm going to play is a key account manager in the tech automation space. So I need to go and deal with lots of key CXOs, and I need to have more than technical skills; I need to have good people skills, good social skills. These are not essential criteria, but that will help me a lot in being successful in my role so that I really like it.

TIM: Yeah, and that's absolutely fair enough. Ultimately, if you're in consulting, you've got to be likable, and you've got to be able to get along with people. I feel like that's essential, and I could almost just be part of the explicit criteria, like knowledge of this, knowledge of that, this type of skill, likeability—you could almost incorporate that into framework I feel Oh, what about this? I tend to hear a lot of hiring managers almost slightly dismiss the technical skills a little bit and say soft skills are more important. You can't really change your soft skills. I can teach anyone technical skills, which I get at a certain level. and we've discussed this a little bit, but can that be overdone? You're not going to get a fresh grad who's got the best soft skills but no technical knowledge and teach them everything. It took me years just to learn SQL, and that's a reasonably simple language to learn to a level that I was effective. I feel like sometimes maybe hiring managers, because they've been doing it for so long, forget how hard it was to really become an expert in their foundational technical skill set, so can you ever remember any examples of where you've seen someone get hired who's been like a great cultural fit and has had those strong communication skills but just couldn't do the job? They didn't have the fundamental technical skills.

PRAVEEN: If I remember correctly, yeah, so from my real-life scenarios, but this is a problem that happens, right? So these are the real-life problems that hiring managers and, most of the time, senior leadership teams go through because we make an adjustment. Most of the time it's a subjective judgment when we select or reject a candidate, right? There is nothing called okay, this is not a university exam. Here is the paper; go and come up with the answer right, so there is a possibility that, okay, you have made a mistake. The person was very good, very nice, a very good team fit, a cultural fit, excellent communication skills, and very good social skills, but they start me with the technical skill set. So we hired the person in the wrong role, so some soul searching needed to be done by the hiring manager: Hey, what's the purpose of the role, and what kind of skill set are you looking for it? So is it the best combination, or is it you completely hiring a project manager into a data analyst role that will not work right? that is bound to fail I was supposed to become a hockey player, but if I go in and start playing football or I start singing, I'm absolutely miserable. I will feel badly, right? You know that. Okay, what specific role are you hiring for, and what technical skills are you looking for? I would say, yeah, it's still paramount. There is no and those people who believe that okay, technical skills can be learned easily. It's not. If it is, let's say just for an example, if it is 120 or 130 Python line codes, and if you have to fix two problems, it takes a minimum of one hour to a very good Python developer, okay? Very good, and if it is average, it might take three or four hours, and if it is 1000 lines, it will take an entire day, so it's not easy, and more than learning ability, learning temperament, and learning patience to learn those skills is also important, which cannot be measured right. which cannot be objectively measured. It can be assessed highly, but a high level, but you cannot assess a person who is very good with other skill sets, but technically skill sets can be taught and learned, but how would you rate that? Okay, the person is having the right temperament to learn the technical stuff or the technical coding. Coding is boring, right? Let's be honest: coding is not fun. Some people say, Okay, it's fun, but the majority of the again 80/20, I do coding. I'm very good with coding. I'm doing it myself. I've been doing coding for the last 24 years, but do I really enjoy it when I have to do coding? No, because if I have to do it, I will do it. But some people do really enjoy it, and nobody likes to spend three hours or five hours just fixing one logical statement that went wrong or some error, so we need to be honest with ourselves, and yeah, so that's what I'm thinking, and one person I do remember that we have hired in us in Citibank, the person was very good, very nice. both techniques This is a third dimension of the problem, not the second one. One person was very good with technical skillset and very good with the social skillset communication or project management, but the person was having a lot of issues with other issues like work ethics, which I don't want to highlight right now. but the person was never on time in the office, right, and always misses the important meetings, and once, twice, five times, we all understand there might be some urgency, but if it happens 10 times, 20 times, then you realize, okay, there is some serious problem, so yeah, so it's not only the technical skill set; personal skill set also, though I don't know how we can measure or assess the work ethics of a candidate. but it plays a major role

TIM: Yeah, that's a great thing to highlight, and as you've just referred to it as a third dimension, that's a really nice way to put it and something that Yeah, it's definitely not measured at all in the hiring process, probably because it's impossible or hard. I guess you could ask candidates to provide evidence of where they've worked really hard, where they've, I don't know, done a degree, and they've worked at the same time. Or they've raised a kid and had a job at the same time, or something like that, but then it's, yeah, it's the kind of thing you only really see once they join, isn't it? And you're right. I don't think we talk about it enough, and it makes the difference between someone potentially being great and very mediocre.

PRAVEEN: and I think it's a good problem for probably the, what do you call it, the psychologist or those who are good at framing the questions for the recruitment from the HR side. These are my wild suggestions, right? If they can come up with some kind of assessment, hey, can we measure it? Can we base it on the responses to questions, right? It's like a psychometric test, right? Or something like that, which helps us to understand, okay, this person may not be excellent in communication, not very excellent in technical skills, but has a very strong work ethic. Okay, probably I need that kind of guy, and we all know that, okay, this work ethic problem does exist in our team in real life, and we all go through it, right? But we cannot ignore it. so some kind of psychometric assessment some kind of questionnaire format that gives some insight right some kind of sneak peek into the candidate, okay, the person is having very strong work ethics and may not be on the scale of one to 10 in terms of communication; it's six or seven. Yeah, my work will be done fine. I'm okay with it. He or she can articulate the technical problem in simple language to stakeholders. A person may not be the excellent world's top 2 percent of data scientists or data engineers, but it can get the work done, and having being strong work ethics is important If he or she needs help, he or she can raise a hand like I did. Praveen I need help. I'm not understanding this thing, or I need to ask him for help. It's not a matter of shame; it's a good thing. We should appreciate this thing, right?

TIM: Yeah, thinking about it now because you mentioned the psychological tests Yeah, the way psychologists would try to measure this is actually, I think, through the big five personality model, the OCEAN model, so one of the traits is conscientiousness. I guess the challenge is that they're surveys, so I would be able to fill out that survey and make the answer whatever I want it to be. so I could appear to be very conscientious if I was cunning by doing the correct answer on the survey, but that doesn't really give you the truth, does it? And it's just once you see them, once you work with them day to day, you're like, Oh, okay, now I understand, and you start to see all those little things, like being late to meetings or not getting to work on time or turning up late during the day, it's only once you see that firsthand that you realize there's a problem as a final quick question I'm wondering if you have had anyone in your career who you've learned a lot about hiring from, like anyone who's a little bit inspirational in their approach to hiring, or maybe they did it in a little bit different way or a creative way—perhaps someone that hired you and whose approach really resonated with you.

PRAVEEN: Yes, this is a very good example. One was in my very early career when I started my job as a trainee software engineer. The recruitment process, I really enjoyed it, right? It was like a simple written test, right? And they said the written test is not a pass/fail test. so you don't have to cross a cutoff or threshold, then after that it is followed by a normal What do you call it? A small group discussion, and again it was not pass and fail, and then the third was it's a staff engagement exercise, so I don't know if it's something which recently I heard in civil services they give you the opportunity to go and talk to existing employees and showcase how good you are as a team player, how you mingle with the team, and how you listen. In communication, active listening is also part of the communication skill, right? And especially from where I come from in Asia, right? Indian culture has an immense emphasis on speaking, right? No, it's not that the communication skill comprises both listening and speaking; right, active listening So that gives me a very good opportunity, and a staff engagement exercise was very good, right? So where we get to know about the various roles, because as a training engineer, I didn't know anything about the organization structure, project manager, scrum master, or business analyst, so that was very good, and that gave me a lot of motivation and excitement to join this team. So I would say this has an element of engagement. Mix and match right again; there's no perfect solution. Having a traditional Q&A format of an interview is good, but how about engaging the candidate proactively with informal chat, allowing the candidate to respond to a couple of critical questions? real-life scenario questions about their approach to video interaction or recorded video interviews And also the recruitment webinar, where candidates can engage right with the team, so these are the, I would say, good practices, best practices, which will help to assist the candidate in different dimensions and in a different perspective and also an opportunity for a candidate, potential candidate, to look into the team and the hiring manager and the role in more detail in a different perspective. So the combination of all these things and also the focus should be I do remember in my first example, which I was giving, they didn't ask any coding problem to fix the problem. In those days, Java was a very popular language, so we all learned Java and wrote a lot, but they gave us a simple algorithm, a simple real-life problem, which was having a closed loop with the code going back to the infinite loop. and they were—they just asked Hey, can you analyze? It's a very simple six-step flow. What is the problem? So with any common sense, even if you're not a programmer with common logic, we'll say, Okay, the loop is not getting close; it's going back again and again. The program will never terminate, right? So focusing on the real-life problems or doing an approach along with tests, etc., made that interview really good. I would say

TIM: awesome