Episode 115

Nurit Lamy on The Human Element in AI-Assisted Hiring: Balancing Technology and Personal Insight

TIM: We are live on the Objective Hiring Show with Nurit. Welcome to the show. Thank you so much for joining us.

NURIT: Thank you for having me. Happy to be here.

TIM: It is absolutely our pleasure to have you here. And I'd love to start by hearing just a little bit more about yourself. Who are you, and who are we listening to today, just so our audience can start to learn a little bit more about you?

NURIT: Okay, so I'm the rate. I'm 42 years old. Two days ago.

TIM: Happy birthday.

NURIT: Thank you. I live in Israel, and I am married. I have two kids, a boy and a girl. I've been with DataWorlds for 17 years, and as a manager, about 10 or 11 years out of that. Hiring data people is very close to my heart.

TIM: And I noticed from your LinkedIn profile, you've got a quote there. It says without data, you're just another person with an opinion. I've noticed in hiring, it's actually not that common. I don't think they really use a lot of data in making the hiring decisions. I find that most companies and most people would rely on something very intuitive. Gut-feel kind of process. What do you think about that? Do you try to take a more data-driven hiring process? Is there some reason why we would not use data in hiring given how important a decision it is?

NURIT: So I think hiring has to complete tracks. We have the very professional side of hiring that you have to have certain skills, and that can be very data-driven. It's even binary. It's either that, or you don't know, and/or you need to know something like SQL at a certain level, but there is the very personal, intuitive part of hiring. Usually they call it culture feet. And that is something that is very hard to put in a frame and be data-driven because it's more based on the connection you have with the candidate and the connection the candidate has with the hiring manager and how you think he will be a part of your team and can really be integrated within your team. And you want to have a people person, for example, with data, because an analyst has many stakeholders. And that's something that is harder to have to be. Data-driven, because you can say, Okay, nice checks. I don't know; courtesy checked. It's harder to measure.

TIM: I agree. Although I do feel like we could probably apply some of the same methodology to these more subjective cultural fit things. Like, for example, you. I actually wouldn't have a problem with a company saying, you know what, 20 percent of the criteria for this role is how likable the candidate is. So everyone who interviews them is going to say how much they like them on a scale of 1 to 10. At least then that's adding some kind of metrics as opposed to some general feeling of I didn't really get along with them. You know what I mean? Would that be a better way to do it?

NURIT: Yes. So what I got to do in larger companies that have things like company values. So we usually say these are, we have, for example, we have six values. These are the top three that are must-haves for a candidate. And then we measure by that. We want to see that, and it usually helps. It's like you said, so how much do you think on a scale of one to 10 this person represents these values that we chose? Usually we choose the values that are really more like a personal thing, a character thing, and then, yeah, and then we rate this. Because I agree. It's very personal. You can even sometimes feel unlucky because a developer A interviewed you instead of developer B, and the entire interview could have gone in a different way. Yeah, I completely agree. We try to quantify this with values.

TIM: And I'm interested in this: have you taken it in the past to a level where it is, at the end of the day, purely numbers? You measure the hard skills; you measure the values. At the end of the day, a candidate's got a score out of a hundred. This one scored 90, while this one scored 80. Therefore, we're hiring the one who scored 90. Or is it, is there still some kind of subjective final decision element that really isn't encapsulated in the number?

NURIT: So usually, yes, hopefully you get to a running of two good candidates, and you try to compare. Okay. So first on the really necessary skills, the technical skills, you try to compare who is better in the technical skills, but let's say they're really neck and neck. Yeah, you're like, okay, let's look at the rating of the values. And you try to take the person that had the highest score or fit. Again, as data people, yeah, I try to make it very data-driven and not based on a personality because it really depends on who the interviewers were. So I try to take that out of the equation and really focus on numbers.

TIM: Yes, I had an interesting conversation this morning with someone that I'd like to get your thoughts on. So they had described their whole hiring philosophy over 20 years as basically trying to find candidates who are systematically undervalued by the market in some way. And so what they noticed was that there are certain types of personalities and people who are going to interview, not very well, like maybe they're not super extroverted. They're not super confident. Maybe they're not people; maybe they're sometimes a little bit on the spectrum. And so he made a habit of just identifying these people who just don't get the jobs they should based on their skills and giving them an opportunity that other people had not given them. What do you think of that philosophy? Do you think there's some kind of merit to that?

NURIT: First of all, it's very nice to hear that someone actually paid attention. And I do agree, and I've had my share of interviews, and I've noticed candidates that had a bad interview, but I could see that there is a potential there. But I know interview anxiety. I have friends who are stuck in jobs just because they're so anxious of getting an interview. On the other hand, in data. It's most of all just people that I ask my colleagues, Do you know people? And then it helps me with their recommendations because I know, okay, even if the candidate had a bad interview, they come with great references. So I trust the references. I know you can have a bad day. You can be very excited in an interview. So I do try to look past that.

TIM: You mentioned something in passing that I was struck by. You said you have some friends who are stuck in jobs they don't like because they have a fear of interviewing. Is that the case? That's really interesting that it's that much of a fear sometimes that we might not even pursue the job that we deserve or we really want. I'm wondering, do you have any advice for people like that, who might have that kind of worry or concern?

NURIT: First of all, yes, I didn't know it existed. I agree. Everyone gets excited and a bit nervous, but I have a really good friend. She's a super developer, but she's been stuck for five years. In this job, she could get a lot more money. I can guarantee that. And I actually referenced her in, like, for friends in a few companies, but she was really anxious, even at a level to start crying during an interview. Like literally panic. attack. Actually, what I recommend I did this with her also, which is really maybe to take a mentor and really do a demo interview. Zoom, face to face, look for questions online. You have many exams, tests, and examples, but I think the best thing is to really have a demo interview with someone in your field. There are many in Israel, many mentoring programs, many people who volunteer, and even people who pay someone, but I think the more you practice, the more you learn. And I think with the stranger, it is even better because it really gives you the feel of a true interview. And I think it's sad because they can go through a hiring process, CV screening, and initial phone call, but then they get to the actual interview and, like, their great first impression just goes to waste.

TIM: Yes. Yes. The difference between interviewing in the comfort of your bedroom or home office versus being there in person in this strange environment, which, again, from a conversation I had this morning, someone described as often. The interview rooms are often very sterile, almost sometimes reminding them of a doctor's office where not that many people are going to be that happy. And so that can be a difficult transition. I think even for me, because I've been working remotely for six years, if I suddenly had to have an interview in an office with all these people that I didn't know, that would be like a little bit of an anxiety provoker for sure. I guess it's exposure therapy, isn't it? That any psychologist would say, just voluntarily, little by little, expose yourself to the thing that you're afraid of. I guess it's the same thing here.

NURIT: Yeah, I agree. And I think also specifically in high tech, most of the offices are very transparent, and the meeting rooms are all made of glass. So you even feel like you're sitting in an aquarium; everyone knows that you're in an interview, whoever passes through the room and looks at you, and then they give you some exercise that you need to solve on the board. Yeah. It's, I think it's that feeling of, okay, they know I'm in an interview; everyone who's passing by the room can see me. I think this is something that really turns the anxiety even higher. I always try to interview in a side room or something that doesn't have so many people passing by to help maybe the candidates, because I know that's not the best feeling.

TIM: I remember years ago actually seeing a really interesting analysis that Canva, the Australian tech company, put out where they were trying to optimize their hiring of software engineers. Specifically, they noticed there was a huge drop-off at one point in their funnel where I think they were giving candidates either a live coding challenge, like a whiteboarding kind of challenge, or a take-home case study. I can't remember which one. They had a big drop-off. They made a really interesting switch in that process. They said at that stage, the candidate gets to choose. You can either do a 30-minute whiteboarding session or you could do a take-home case study. And so what they found was that the process was accidentally selecting for either, I can't remember whether it was like extroverted people who wanted to do an interview or introverts who wanted to do the take home. And by giving the option to candidates, suddenly they got a better balance of both. And their drop-off actually went to 0 percent in that stage. which I found really interesting. This was years ago now. I'm sure they've changed it. It's what I'm always struck by in hiring: how much of it is very biased or beneficial or whatever disadvantages certain segments and not others. And interviews are one of those, which surely extroverted, very confident-sounding people. Who are beautiful, have an advantage over others, and in a way that it's not necessarily even correlated with the job, that's a

NURIT: I agree the processes are set for a long time. It's been okay with the CV screening and then an initial phone call and then the hiring manager and then a home assignment or something. And it's very sad. It hasn't changed for a very long time. I've been in high tech for 17 years. I've never seen a process that was very different, but maybe one last interview or you don't have the home assignment or you do have the home assignment. It's not very different. So right now I was actually sitting with the talent acquisition that's responsible for my team. And trying to think outside the box because it's so hard to find good analysts today. But the market is flooded with people with an analyst title, so trying to think outside the box, how can we match the process to what we're looking for? Like you said, to overcome all these obstacles of getting the right people or giving people a really equal opportunity, D and I's are always spoken, but I agree. It's not always something that you can see, like, how do I make the process fit people with anxiety from interviews? Like you said, this can fit extroverts, and this cannot fit. Really trying to have in my arsenal, I don't know, two or three types of processes and try to match. Because when you just read the CV, you don't know what will match the person the best, but maybe try to understand in the first phone call what will fit. So we still don't have the, like, the perfect answer, but we're definitely trying to have these few options.

TIM: Yes. And one of the hardest things I think in hiring is, I feel like it's a game of trade-offs where In this example, maybe you make the process a little bit more flexible to accommodate different types of candidates, but then it becomes maybe a bit more complicated, maybe harder to compare. If you had, I don't know, in the case of Canva, some people doing a whiteboarding session, some people doing a take home case study. It's not obvious how you compare the results of the candidates, so that must be a challenge as well,

NURIT: It's a challenge. Yes, because it is. You don't necessarily have the common ground that you can compare. And I think also a concern that I recently had is when I give a home assignment. I'm not sure that the candidate actually was the one who solved it, especially with AI tools today. And this is why if the assignment was. was good, then we have an interview on the home assignment where we ask the questions to make sure it's not a bit about the coding. It's about the way of thinking specifically for analysts. I want to see that this analytical thinking is yours, like the candidates and not some neighbor, someone they paid for an AI tool. GPT can very easily write SQL queries. You don't get the analytical thinking. So it's another aspect that makes it harder to compare that you have to be sure. It's a whiteboard. It's the person, but it's a home assignment. Then now in the AI age, we have more, even more, concerns about it.

TIM: Yes, but as you say, with the follow-up interview to delve into the details, I would have thought It would be very hard for a candidate who had either outsourced the entire thing to their friend or to Chachapiti and had just sent it off that they would really drown in that interview pretty quickly if they didn't understand what they'd done because you, as an expert interviewer, could really start digging into the details.

NURIT: Yeah, you'd be surprised. I already had the case of a person; for example, we asked something about employees. The queries were about employees and, but in the, during the interview, sometimes he said employees, sometimes he used other terms and it as, as more and more that I asked, it became very clear that he was not the one. Who solved this exercise? I think it was ready with answers on the very top surface. It was not ready for me to really deep dive and ask really harder questions, I am happy that I was able to find out that, before I hired this person, because really the exercise was really well. So I hope he learned his lesson, but I definitely learned my lesson that I have to deep dive if I want to get the hiring to get the good candidates. And if I wanted this hiring process to be successful, it even made me consider not giving home assignments because of that. Okay.

TIM: Yes, it's, yeah, AI has just changed so many things, and I feel like we're in this weird state where candidates have adopted it very quickly for their resumes, sometimes in interviews, sometimes for tests. And companies haven't quite caught up yet because the software hasn't necessarily been built out using the AI. We're in this weird pattern where the hiring process seems a little bit broken. Like candidates are using it to write CVs, candidates are using it to apply also to jobs. And so maybe we just need to see this period out; maybe in a year or so, the HR tech might have caught up to have a sort of AI-versus-AI battle. I don't know. Where do you see this going in the next couple of years?

NURIT: I definitely see AI going into the screening process and the hiring process. Today we have these automatic algorithms that can screen CVS, but I already know that they're not always the best. Even when I was looking for a job, I got this automatic no, that I am not a good candidate for this job. And then my friend working at the same company went with my CV. The talent acquisition showed her my CV, and she was like, Great. Can you submit her CV? And she's my friend, said she just got a no from this automatic algorithm. And the only reason was because I didn't emphasize the specific term that the algorithm was looking for. So I think we have a long way to go with AI. I think it started in HR tech, but it's definitely not at the same level that candidates can already use it. So I think the AI on the hiring side needs to be more mature.

TIM: Yes, for sure. And I assume we'll see emerge in the next year or so the next greenhouse lever, smart recruiters, or whatever that new wave of HR tech is that's AI native that will hopefully come along and solve some of these problems. But I feel like even with, let's say, the most accurate AI CV screening tool, as an example, the problem is the data is crap. If you're trying to compare a resume to a job ad and the job ad is not that reflective of the job and the resume is not that reflective of the person. It doesn't matter how well you match those data sets. It's like voice. It's still going to end up with a lousy business result. Do we need almost some better quality data about the candidate and the job? And then whatever tool we use to match it actually has a chance of succeeding.

NURIT: I think this is the reason why. Some people say, I'm not afraid AI will replace me because I think specifically in hiring, especially in hiring, there will always be a room for the human factor as good as AI will be and algorithms and comparisons. And I don't know what you will always have to have this human factor talking to a person to person. Understanding what's going on, and you have to be on the data, right? Let's say you have, we have this automatic process. We're not interfering, right? But then you measure your employee attrition or involuntary attrition. So you have to see out of a hundred that were hired with this automation. No, no human interference. How many were good fit? In terms of a year, how many were laid off? So I think as good as AI will get in hiring, we will always have to have the human factor to actually see this match, see a fit, even not, let's say AI could do the technical stuff, but on the personal level, on the culture feed, I don't know, team player, you'll have to have a person.

TIM: Yes, for sure. And you were talking about almost like comparing the success level of AI versus human recruitment, let's say, I think. Part of the issue there is, I don't think we even have agreement yet on what success is because you could measure it as, yeah, stays in the company for a year. Is that successful or not? What happens if they leave after a year to get promoted elsewhere? You could argue that's successful. What happens if they stay in the same job for five years in the company? Maybe that's unsuccessful because they haven't advanced their career. So we'd have to get some measure of success to begin with, which I don't think a lot of companies even have.

NURIT: Yeah, I agree. I think today companies measure it in very black-and-white zeros and ones. Did you stay, or did you leave, and how long did it take you? Okay. Let's say benchmarks are. 7 percent in the industry, and we have a 5 percent attrition. So we are in a very good place. And we are in a good place, but I agree that it's not that black and white, because, like you said, maybe I'm happy at my job, but I just want to get promoted. And the company that I work at doesn't have this opportunity, not because they don't want to promote me. Or they don't value me. They just don't have this opportunity. It could be 10 people start up, and I want to be a team lead, but maybe it will happen in three years, but not now. And somewhere else offers me this team lead. So I might go. So I agree. It is again the human factor. I think we'll have to have a, like, a sentiment analysis, maybe ask people, see why they're leaving. If it's the company or something internal, or if it's just because they had a better opportunity, I agree. I know people who are in my teams, and they're still in the same role. They're not, but they're not looking to move forward. They are moving forward with their ranks within the corporate, but they're not looking to be managers or tech leads or something, but they're good with it. So again, is this a good success criteria? Because they're essentially not moving, right? But for them, it's what they want. So we'll have to have these, I don't know, like surveys. Or something like that, that will give us a more complete picture of this. But we'll have to get sentiment analysis into the process.

TIM: Yes, maybe the performance reviews as well, which maybe add some value, but then also you could say sometimes a bit biased. I'm sure people listening might've had a performance review they didn't agree with. So yeah, it's a complicated problem. Is there any step of the recruitment process that you think we should just throw in the bin? Like, it's just, it's worth pretty much nothing. It's not worth our time doing.

NURIT: I think we can definitely maybe skip some levels or maybe maybe combine some things. I think it's redundant to have, I do the CV screening, and then sometimes HR in many companies, HR does these first initial calls of 10, 15 minutes to tell about the company and ask what the candidate is looking for, like certain words. And then sometimes the hiring manager does this first interview or something. So I think, especially in the beginning, we should have the process a little shorter. Maybe have the hiring manager CV screen. I know it's a lot more work, but I think it will be more efficient, especially in a data world. If I were to do the CV screening and then these first initial calls. I will make the process shorter, and I will not waste HR time and candidate time because I know what I'm looking for. I can ask better professional questions to screen people. Again, not that they chart, they know the job, they know the roles, they know what to do, but I think a hiring manager. Those in this field can understand the answers better, and they know what they're looking for, so they can ask a question on top of a question to get a better understanding, and then the candidate either rules out or goes to the next level. So I think we can save some time there.

TIM: Yes. It must be so tricky for people in talent acquisition and HR to do the initial recruitment for roles that are so far away from their own skill set or experience. Like, I try to put myself in their shoes. If I were trying to hire a lawyer or, I don't know, a civil engineer, I would be clutching at straws. The best I could do is ask the hiring manager and try to get a sense from them. And talk to them about what I'm looking for, but it's never going to meet the quality of an expert doing the screening. But then the opportunity cost is your time. of

NURIT: Yeah, I definitely agree. They have a very hard job, and sometimes they're responsible for 20 open positions, and also it's very hard to do this context switch. Okay. This candidate that I'm going to call now, is he a developer? Is he a data person? Is he a product manager? It is very hard to have this context switch, I think, all the time. And like you said, it's very far from if I need to interview now a developer, and I work with developers and everything, but I'm not, I don't know, they will tell me I use. React. Okay. That tells me nothing, really. So yeah, I agree. It's very hard for them, and as much as they will sit with the hiring manager to get whatever tips and tricks and what are you looking for, I think it's very hard for them. And I think if a hiring manager has the time, or more correctly, they need to make sure they have the time to do this. And I think give HR their specialty to interview the candidates really on the HR aspect and not waste your time doing something that the hiring manager can do faster and more efficiently and have the entire process. I think, HR, they can call 50 candidates, and I would rule out 25 of them in the CV screening process.

TIM: One other benefit, I think, of those first stages being done by the hiring manager is, at least in my experience, that's who the candidate wants to speak to as soon as possible. Because that's their manager or their manager's manager or whatever. They're the ones who actually understand the details, the expectations, the goals. They're the ones who can normally answer their questions in the way that a normal external recruiter, HR, or talent partner would struggle to do. And so from the candidate's perspective, they'd normally be pretty happy with that, I think, as well.

NURIT: I definitely agree with that. I think. Even when I was looking for a job and to get that first phone call, and it's someone that is not on the technical side, and I ask questions like, Which tools are you using? Are you using this? Are you using that? What are the databases? What's the visualization? And yeah, the people in Talent Acquisition, they don't need to have these answers, right? So I agree it can save time for all parties involved, and yeah, it will give HR to have, and I think it's exactly like in our day-to-day job, you know? Each of us has our own expertise, so we should use them in the hiring process as well. If I'm the data person, I'll talk; I'll screen the candidates on the data side, and HR should screen on the HR side.

TIM: I'm sure you must have read many hundreds of resumes, probably thousands. Actually, have you noticed any common patterns where candidates fall down in the way they've created their resumes, the way they've crafted them, or any kind of suggestions you might have for them?

NURIT: Yes, I think People sometimes elaborate where they shouldn't, and they don't elaborate where they should. I think a candidate should be really mindful of the roles that they are applying for. You should even have several CV types. For each role that you are trying to apply for. And if you're going for a data analyst, elaborate on the things that you did in your previous roles that are very much related. If you started as a product manager and then went to being an analyst. Okay, write it like a row or two that you are a product manager, but then give me like the five or six rows about being an analyst and the tools you worked with and some interesting insights that you may have, because I really see this thing as a mistake. I can say, repeating that even sometimes I get a resume of just, I don't know, I worked in this company from 2010 to 2015. And like a title data analyst and then nothing else. So for me, it's okay, but what did you do? I'm not going to call you and have an hour-long phone call. Just so, tell me what you did. For me, that's maybe a miss, right? Maybe this person is a great analyst, but there is nothing written under this. So I think you should be very mindful of the position that you're about to apply for and make sure your resume really reflects your experience that is related to this. And if you need to have three different versions of your resume, so be it.

TIM: A bit more; are they starting to do better in that they seem to match the job description better?

NURIT: So I do feel there are trends. In the way CVs look like every few years there is this new trend now It's two columns. Now you have a colored column on the left and things like that. I think it's a trend thing, but definitely I feel that they all have the same kind of patterns, and they're definitely trying to think about what the LLM will be looking for and then try to match that. And actually, this is a tip that I give. I do a lot of mentoring for junior analysts, and sometimes I tell them. Go on LinkedIn, go on all these places that jobs are posted, and see what they're looking for, and make sure you have these words in your CV. So if they're looking, someone that has worked with, I don't know, snowflake and you use snowflake, you have that experience. So make sure that you write this down. It won't necessarily get you through the CV screening, but it can definitely help. Or make sure it's in a certain type of file. For a PDF or Word, make sure that when the algorithm goes through this, it's not gibberish; it can actually read English. So definitely, I think people are trying to think about what will pass these LLMs, like trying to find these keywords to

TIM: Yes. Yes. And I think one of the challenges has always been that, and maybe this is slightly subtle, but when you apply for a job with a resume, you have very different audiences who are going to read that resume. You have, let's say, a recruiter, talent acquisition, HR person, and hiring manager person, very different personas in terms of their knowledge probably of what you're talking about on the CV. So it's almost like it has to be pitched at different levels, even though it's one document. And now, as you say, there's this third audience, the large language model, which will start to become more and more common. So it reminds me a little bit of Google search engine optimization. You've got to somehow create a website that suits humans and the Google crawl bot. And I guess we're going to go down the same path with resumes as well.

NURIT: Yeah, I agree. I think this is what kind of the look-alike, the look and feel of the CV, is trying to do. Okay, not to have too many words, but not too few words is for the person that's just going through the CVs with their eyes, but not actually reading what's written, and maybe put some words in bold. So that's more, I think, for the maybe talent acquisition that is trying to screen according to what the technical hiring manager said. So they will be putting in bold their, I don't know, their SQL or analysis or words like that if it's for analyst positions. And then I think the actual content, like the elaboration of each role, is for the eyes of the more hiring manager technical person. So I think this is the way to try and make everyone happy. When they're looking at this one page, hopefully one page, because there's also something that says don't do more than one page, but then what if I have a lot of experience? So the last tip that I got was to summarize. I don't know from the, if I 17 years, so I don't know, summarize the first five to 10 years in different data roles in the following companies, but don't elaborate, just elaborate on the latest. So I think it's also a trend, and it really depends on the person giving you the advice.

TIM: Yes, it's so subjective. And what. I imagine it is going to happen. This would be my guess: having the idea of having this short resume has been because the people who have been reading it are the ones who don't want to scan over 100 pages of the document and get lost in it. Like they've only got a minute to look at it. So it's needed to be two pages long. The problem with that is it's not really that much data to make a decision. Maybe now if LLMs are going to start doing the screen, maybe there'll be this. elaborated full CV that you submit. That's like everything you can think of about your work experience, all the data, and maybe even rich video content of conferences you've attended. It could be anything, but then there's some kind of human summary, either through the system, which adds an AI summary, or you add your own. So that you could almost have, yeah, the full data set to appeal to the AI and then the truncated one for the humans. Will it go down that route, do you think?

NURIT: I think definitely more and more automation will go into the process. I think we should get more creative. I think if people had the opportunity to have these 30 seconds of self-pitch to go along with their application, it may help. I'm not sure everyone will do it, but I think it could definitely Screening better because I don't know if I will say, this, I have, this is my CV and I go and I have a 30 minute I don't know, video that I say, hi, I'm the way that I have. The many years of experience in data roles, I don't know. I think it has a beneficial advantage, and I think it will be necessary because all these white pages with words on them, it's very hard to differentiate. So I think the more automation and less, like you said, people don't have to sit in front of the computer and do this, we can afford to have to screen a lot more CVs in less time. So I think we have, we will have to go there in order to really bring quality talent. And I don't know, I think companies will want that also to give them the Advantage on other companies in bringing good talent

TIM: I feel like at the moment we're in this situation where the inbound flow of candidates through job boards is a little bit messed up because of the AI-generated CVs and the automatic applications and what have you. But as you mentioned before, referrals are still a really good way to identify a good candidate, especially if the referral is from a trustworthy source. Would you focus, if you were a candidate, on trying to get a referral into a company, rather than applying through jobs boards and fighting it out with another thousand applicants?

NURIT: Definitely through referrals. It's the top hiring channel. Because I think the company said we have very good talent. We trust our employees. So if my employee's trust brings me a referral, then I definitely trust this referral more than just some random CVS. And I think this is why in most referral forms, we also have, How well do you know this person? right? Because I get a lot of LinkedIn messages. Okay. I see you work at High Bob. Can you please submit my CV to this position? So I do put the disclaimer, and I'm being super honest. I don't know this person. They. Reached out on LinkedIn, but if this is someone that I really know, then definitely I will give them the very warm recommendation and say that I know of this person I've worked with this person, I can vouch for them. And when talent acquisition sees such a referral, then they definitely prioritize these versus something that comes from LinkedIn or, I don't know, some other source. Job board. And I think it makes sense. Again, I know this employee. I've hired this employee. So if I want them, I want their friends, specifically if they refer them to me. And again, like in the data world, it's very hard to find good analysts. I will always turn first to my friends. Sometimes, my friends, they interview people. And like we spoke in the beginning, they had two good candidates, but only one position. So they chose so, but they will recommend the other person that they sadly couldn't hire. So I would do a reach out to these people, and I definitely trust these more than just random CVs. And I, maybe I'm missing something. Yeah. I could be missing the best analysts. But still. We trust the people that we trust.

TIM: So it's almost a case of managing risk on some level because you're hiring someone, or. Someone you know, who knows them, reduces that risk of a regretted hire quite a lot. You know what you're going to get most of the time. Is that fair to say?

NURIT: Yes, definitely. You will prefer hiring someone that comes very recommended. It could fail, right? I actually had this case that a friend of mine was looking to hire someone that was someone that used to be on my team in one company that was, I knew he was looking for a job. He got a very warm recommendation from me and also from the manager who managed me and managed him after me. But it ended up not being a culture fit. He was laid off after three months. And my friend said, how could you recommend this guy? And I'm like, listen, when I was his manager for five years and he was like one of my top employees, I don't know what changed in the middle, but I was managing him in. And she said, No, the manager after you also recommended him. So I think it could be sometimes a culture fit, maybe a fit with the manager, or sometimes something just doesn't work. Connect, right? And he came super recommended. So that can happen, but I think in the 95-plus percent, it will be a better match than someone that came from a job board. It will be more successful.

TIM: If you had the proverbial magic wand, and you could wave the wand and fix the hiring process, how would you wave it? What would that perfect hiring process look like?

NURIT: I think I will have some automated model, but something that I can teach, with natural language processes and things like that. To have this very tailored per position, this is like magic, not something that just screens for everything. Like I would have tailored processes for analysts, developers, and product managers, obviously other sectors, not just high tech. So that's the magic I would want.

TIM: I like the sound of that magic, and maybe it's not too far away. LLMs are advancing at a breathtaking pace, right? So fingers crossed. If you could ask our next guest one question about hiring, what would you ask them?

NURIT: Maybe they know good analysts, but no, I would actually ask if they also have good tips for an accurate hiring because we all only know the people that we've met during our careers, and we speak to colleagues or meetups and conferences, but it's always good to ask more people about their tips for a more accurate hiring. Okay, yes. Accuracy is so important, especially because anyone who's ever made a regretted hire, I think, will never forget that because it lives long in the memory, and it's a much worse problem than the ones that got away, like the false positives. burn more than the false negatives, which you don't really notice.

TIM: So hiring accuracy is always top of mind. I know when I've seen academic studies of this, like what actually predicts job performance, it's quite interesting to me that it doesn't really connect that closely with what companies actually do. So the top predictors on average the intelligence of the person. their conscientiousness, like using the big five personality model. And then a third place would be like job skills tests. But a lot of the things that aren't predictive are other things we spend most of our time doing. How much experience do you have? The resume and unstructured interviews rank really low. At least based on the academic research that I've seen. So it's, yeah, I've always found that interesting that there's this weird disconnect between practice and academia.

NURIT: Yeah, on one hand, we are trying to find this formula that will match or candidates and all the hiring processes. But at the end of the day, it's not just that each role is different. Each person is different. You can have a master's degree, but you are in this specific field that I'm looking to hire, but you may not have the personality match or, even technical skills; you may not feel completely. I have a master's in computer science. Could I be a developer today? No, because I didn't write any code more than 10 years ago, right? But on paper, I would probably, the LLM will pass my CV, right? Because wow, I have a master's in computer science, but in the first phone call, you'll see I'm not, I'm far from being a good match for that position, right? So I think it's very, we're trying, we need to tailor. This formula is how sometimes the first measure is the most important; sometimes it's the third that becomes the most important. It's about tailoring this.

TIM: Yeah, looking forward to discussing tips for accuracy hiring with the next guest, and in the meantime, it's been a great conversation. Thank you so much for sharing all your thoughts and insights with our audience today.

NURIT: Wow, thank you for having me. This was great. I enjoyed it a lot. I'm looking forward to hearing the next one talking about accuracy.