Episode 138

Luuk Feitsma on Making use of AI and Unconventional Tactics in Recruitment

TIM: We are live on the Objective Hiring Show today. We're joined by Luuk in the Netherlands. Luuk, thank you so much for joining us.

LUUK: Yeah, pleasure.

TIM: And where I always like to start is to hear a little bit more about the guests, just so our audience can start to think about who they're listening to today.

LUUK: Sure. So I'm Luuk. I'm 33, and I've been in recruitment for, I guess, about eight years now. I've worked in the data and AI field for about seven and a half of them and mainly worked at a firm focusing on data and AI and helped grow that team, where I also got taught a lot about data and AI. So when I joined there, they gave me, like, a stack of books this thick to work through. I took a lot of programming courses myself because I had a very, very high curiosity for anything, really. So I just really want to know something about everything. So that's also where I learned how to program Python and SQL and some web development stuff. I took some analytics translation courses, but as a whole, I would say on a functional level, I'm always able to keep up with technical people. But for technical people, if you're ever in need of, like, bad coding examples, then that's what I'm good at because I don't code in my day-to-day.

TIM: That's a great introduction. And you also dabble in a bit of content. You occasionally go viral online. Is that a fair comment?

LUUK: Yeah, you could say that. So I had a big one, of course, last weekend, where I created a bit of fake news. So I thought it'd be funny to do, like, to ridicule the thing we have called Saharan Dust Clouds, and they come across the country, and when in the morning you go to your car. It's covered in dust, and it always seems to happen like within a week of going to a car wash. So I thought it might be funny if I made, like, a video where it looks like drones are dropping the scents in large areas, and it's by order of those car washes to actually get more revenue in because more people need to wash their cars. So I created an AI video that shows a drone dropping 10, and then over the weekend it gained 4 million views.

TIM: And I'm sure you've got something else interesting planned for April Fools. We'll keep an eye out for that. What's your platform of choice? You doing this on, like, Instagram or

LUUK: Oh yeah. So my Instagram is closed off. So that one is really private. But even though I'm 33, I'm still of the generation that quite enjoys TikTok. So that's where I post a lot of my shenanigans, and if something is nice, then I do push it over to LinkedIn. But mostly it's TikTok. Yeah.

TIM: Yeah. Nice. Okay. I'll have to keep an eye out for that. I'll have to log into TikTok for the first time in my life and see some of the shenanigans.

LUUK: There's some good stuff that you can use in there. So I quite often use it when I have an event or a meetup or anything and I need some quick video content around it. If you make, let's say, 20 short videos of five seconds, dump them all into TikTok. Then by using

TIM: Yep. Oh, nice. Yeah. Excellent. Okay. I'll have to check it in and go down that wormhole. I'm always wary of myself with these kinds of technologies, and I could see being in it for eight hours, suddenly realizing, Oh my God, where's my day gone?"

LUUK: I wouldn't even call it a wormhole. I would call it like a black hole of death because you'll end up getting sucked in, and it's hard to get out. Be wary.

TIM: So you've created your little bit of funny fake news, which is great, and although some people are defending its veracity, by the sounds of it. Actually, it reminds me of this. Remember a couple of years ago when someone came out with the idea, the theory, that birds were fake, and they tried to push the fake bird narrative, and then a lot of people actually believed it and are on the street protesting in support of the fact that birds are apparently fake? So you might have created a new movement. I hope not, but we'll see.

LUUK: I know that story, and I know that the people protesting were part of that fake movement. It wasn't as serious, I hope; hopefully, maybe there are a couple of loonies among the normal people there. But yeah, for sure I am a bit worried that it might lead its own life. And I've also made a video explaining how I created it and that it's AI, and I commented to a lot of people, Hey, it's AI; it's not real. But the weird thing is when you, when people believe something is real and you tell them it's not, they seem to double down. It's super hard to get across that. No, it's not real. Be a bit more critical of what you're seeing. It's difficult. I've learned that after that.

TIM: Yes. And from fake news to fake resumes, I guess you could say. A lot of candidates these days are using ChatGPT or Claude to at least optimize their resume based on a job description. But a lot of the sense I've been getting speaking to people is that. If we had some kind of truth index of a resume, it seems to be at an all-time low. Do you, are you seeing the same thing, or is it just people rewording things and making them sound better?

LUUK: No, I think it's a difficult one. But the thing I do see that a lot of people, I think, are making is a summary of your resume using AI and an assistant. In that way, I think it is a good way to go. But the problem is when certain aspects of your work history are being, let's say, exaggerated, like all these new technologies you, they get. But then when you haven't read something. And then AI, of course, doesn't know how much you did or did not do with certain technology, so it might give some weird results, and people do need to be aware of that because, yeah, I've seen it a little bit. It's also that everybody's resume seems to be the exact same way.

TIM: It is. Is it that Euro Pass format? Like I see that seems to be quite common in Europe.

LUUK: Oh, now that one I've seen a lot as well. But no, there are multiple formats. But the way that the sentences are structured, that's mainly the same, right? Because they, they can have maybe some different layouts, but the sentence structure and specific words seem to be a lot more common, like leveraged. I leveraged this to do that. It's something I wouldn't typically see in a typical resume before. It's something that, probably, businesspeople use a lot, and maybe that's the trading data I use.

TIM: Yeah. Yeah. It's a very MBA-businessy term. I'm going to go against the grain, and then I'm going to put levered instead of leverage, just to be a bit of a rascal and see how that goes.

LUUK: Yeah, because I've never seen anyone use that one, so that would stand. out for Sure. And spearheaded also spearheaded that. That's a word that's super common now on resumes since the dawn of AI. Really,

TIM: Yeah. That's so fascinating, the bias in the training data, because if I've ever written the word spearheaded in my life in any context, I don't think so. But now It's on every single person's resume.

LUUK: Some words, you immediately see that this AI-generated text really

TIM: I, the ones I noticed for ChatPilla, ever-evolving landscape. If I ever see those words together, I'm like, come on, man. Just at least Replace that with

LUUK: familiar. Yeah. Also sounds very familiar when.

TIM: And what is your general perception of candidates using. AI in the hiring process. Is it fair game? Is there some limit to what they should and should not be using it for? What? What's your overall view of things at the moment?

LUUK: No, I think it can be super useful, but I think the mistake comes in when you give it too much control and don't check enough. For example, I can see it being very useful when you have a resume already made yourself, and then you find one, two, maybe three companies that you think, oh, this might be interesting for me. And then. With the use of AI, maybe see how you can adjust your resume to stand out more positively because it matches more with what is being sought out there. Because let's say you've done some things in supply chain; this company here focuses on that. So then you highlight more of your relevant experience on that side, right? I think when using it that way, it's super fair game, smart, and saves you a bit of time. I think it becomes a bit more of a mistake when you just let the AI run its course and make decisions really for you about what needs to be on the resume. And it starts hall loosening things. I think especially anything AI-related. When it becomes more towards the decision part of the equation, then I'm a bit wary about it because essentially you're giving out that much responsibility to what essentially boils down to a rock. That we trick it into thinking by shooting lightning into it. And I don't think that's something I'd be comfortable with.

TIM: Yeah, I think it's a very different ball game if, as you say, you've written this comprehensive document and then it's summarizing or rewording based on things as opposed to. The other way around is, Here's the JD. Write the perfect CV for me, without knowing everything I've done. Although I can see why people would be tempted to do that, because that's easier, probably, and quicker and ends up with a more perfect-looking resume for the job description. But I, I think it's such a short-run game for a candidate because, as you say, surely they're going to get into that first interview. You're going to ask them a question about X, and they're like, What is X? It's X on your resume, so you better be able to explain it.

LUUK: Yeah.

TIM: Yeah. One thing I also found slightly odd, I have to say, was that, like the last time we were hiring some salespeople as part of our overall application process, we had a skills test, so fair enough. But we also had just some questions that I wanted to. Know from them. And these were questions that were trying to help make the onboarding process as good as possible. So I was basically asking them things like, Hey, imagine it's your first day at an Uber. What are the three things you would need from us to be as successful in your role as possible? And so it's like, for me, a very personal question. I cared about their exact opinion, wrong or indifferent. I wanted to know what they thought. But the vast majority of candidates just had clearly copied and pasted. From ChatPT, which is not helpful to me because I can ask Chat. Pt, I don't care about the larger language model. I care about what you personally think. And so I feel like that's a misguided use of AI personally.

LUUK: Oh yeah, for sure, because it's something that should be coming from your own opinion. What do you need?

TIM: I did. So most of them I wouldn't have ended up interviewing. But some of them, some of them that I interviewed, I did. I can't exactly remember what they said because it was about a year ago now. What one other candidate I can remember had told me was, Oh, I wanted to get ahead of the game. And so I've developed this sales playbook over the weekend, implying that they put in this extra work, which they didn't need to do at all. But almost hanging their hat on the effort thing and saying, I'm really motivated. Look at what I've done. But then they just copied and pasted something into chat, and Chippy T, I'm like, it's not really of any value, and you are pretending as though you've done it. So it's like a double negative in my view anyway.

LUUK: Yeah, move on.

TIM: I think so. I think so personally. When we chatted, it was probably a month or two ago now. You'd mentioned that you were running some hiring workshops, and the way you were approaching it I thought was really interesting. And I'd love to hear more about what you did in those workshops, why you did them, and how the sessions went.

LUUK: Yeah, sure. So it's a bit of a problem with a lot of hiring processes. The things you're looking for versus what you're interviewing for seem to be a misalignment, and it can be in different aspects, right? Maybe person A has a very different view on what this requirement means than person B. Also, there's some miscommunication between what is asked in different interview stages. Answering the same questions in different stages, which is also, from a candidate's perspective, not a great experience. What I typically do with those hiring sessions or so hiring workshops is that we have these different Post-it colors. Let's say we have three colors: one is hard skills, one is soft skills, and the other one might be culture or educational background, stuff like that. And everybody, there are probably four people present. People who do the role, maybe some sales, and each of them write down for themselves what they feel like a profile for this role is, and one after the other. We start posting it on the wall. And also explain what we mean by something. Because when they write down, they need to be good communicators. And I ask you, and I ask somebody else, what does it mean to be a good communicator? It might be a bit similar, but still it will be two different answers. And we need to attach business-like definitions too. Okay, now we have this huge list of requirements for the profile that we have. So if I find somebody that doesn't have these two things here, are we not going to interview them? Yeah, well, oh, okay. So they're not really required; they're nice to have, and that way you can maybe eliminate some other things there. And then finally, when you have that, you map out all these requirements that you wrote down over the interviewing process. So you have several different steps. What are you asking at which phase of the process, and how are you going to ask it? So that way. What you're asking throughout the whole interview process and then finally to really round things up. You create, like, some example questions for the interview stage with an answer matrix attached to it. So instead of saying, Here's the answer, or Sorry, here's the question and here's the answer we're looking for, we give you an example. Question with some elements of good or bad examples of answers you would typically get. Because there's, of course, a lot more ways to answer a question correctly than just one. Especially in it, I think there's more relative to Rome. And it gives a lot more structure and repeatable outcomes to the interviewing process and has less chance of getting a different outcome depending on the interviewing members you are selecting.

TIM: That is fantastic. And such a great way to solve just so many different problems that you've outlined there in terms of that inconsistency you could get in terms of that subjectivity. Yeah. Great example of the communication skills that you could ask a hundred people what a good communicator is, and I feel like you'd get. Almost a hundred unique answers. Probably like some crossover, but maybe unique,

LUUK: Yeah, but difficult.

TIM: Right?

LUUK: Then, okay, so how do you judge if somebody is good at writing Python? And you ask that, and then you get so many different answers. And that makes it very difficult to make the right decisions when doing interviews and making the decision. This person should head forward towards the next interview stage. 'Cause your opinion on what is good might vary differently from the next person.

TIM: I am wondering how you balanced the. Variety of views you would've had from very different stakeholders. I'm trying to put myself in the shoes of when I've been, let's say, a technical hiring manager in the past. To be honest, I would've thought at the end of the day, my opinion is actually the most important one. Was there some sense of a hierarchy in the different views, or were everyone's views equal? Or how did you, who, who, does someone have like a casting vote? Like, how did that work?

LUUK: Yeah. I'm in the Netherlands, and the Netherlands just is very non-hierarchical, right? We don't like hierarchical situations. And especially for me, I have the mindset that when you are hiring for a team, then that team should have an equal say in who joins that team. So the decision on whether or not somebody progresses through the next stage lies with the people interviewing at that stage and not a hiring manager that wasn't present during that interviewing process. And I think as well, if you, let's say you are an interviewer. And you interview someone together, and of course, there are always two people there, and you both give a negative recommendation, and then the hiring manager says, I don't really care about your opinion. We're still hiring this person. Then that gives such a wrong message about the effort that you put in, because you're also investing time into interviewing somebody and then getting pushed aside. So why would you help out the next time then if your opinion isn't actually valued?

TIM: So if everyone's going to be involved in the hiring process. If interviews are valued, then of course they should be involved at the start as well to define the role.

LUUK: Yeah. And who better to know what's required to be successful in a role than the people already in the role it's going to be? Hire

TIM: Yes.

LUUK: I work in consultancy, right? So there are more people doing that type of role. So we have Aary in.

TIM: Was there anything surprising that came out of this process? Like, I imagine this is something you've done before for other roles as well in other companies. Is there anything surprising that comes out of these sessions?

LUUK: Yeah, so it, it's

TIM: i.

LUUK: mostly to when you are identifying, okay, where are the, where is already alignment on what we're looking for and. It's missing, right? So as you are posting up these post-it notes, you're first doing all the hard skills person by person, so it's only filled by hard skills, and then it's only filled by soft skills, right? You end up clustering all these things together, right? Because maybe the first person, if you're looking for a data engineer, obviously they're going to post Python, and everybody else is going to do it as well. But maybe there's only one person that. Does it have some other technical requirements, right? It starts a discussion. Okay. Why are you posting that? Because I think it's important, because it's used in your work and this and this and this, and it starts this discussion. I think that's the most important bit of doing a workshop like this with the right people present. It also quite often starts a discussion on what the educational background of a certain person should be. Or do we really need somebody with five years of experience? Do we look at what they have done? Actually, oftentimes. Are better at being technical and hands-on than somebody who's done seven years of experience but has done the exact same thing all the time. And the person with three years of experience has a bit more of a T-shaped profile, more brother experience with some specialization there. And the discussions you get there, those are the most valuable you get out of that workshop, really.

TIM: This is great because I feel, as a data person, the way I think about hiring is like a matching problem. You have to have some candidates on one side and a job on the other. But the problem I. That so often the data sets, the data quality on either side of that matching, are pretty crap. Like the average JD that you know you've pulled out of the desk drawer or the fold that you reposted from three years ago is not very accurate. And so going through this session, which might be a fair bit of effort and there are lots of people involved, is well worth it because so many processes, I think, get derailed from step zero because they don't get the definition right in the first place.

LUUK: Yeah. Exactly. And by the way, if we're talking AI hiring, I think. Useful ones I have on that one are that when you look at job descriptions, AI can be super helpful. That's something that I've used for years now. So I'm a big fan, of course, as is everyone, of Jet GPT. And I've had a paid account for years now. I think so. I'm also building a lot of custom GPTs. And something I found very useful is when you fail out, if a custom GPT has a lot of papers about, for example, psychological triggers and inclusive language, but also our corporate. Example projects of things we've done. Our HR documents are saying, like, how, what? Then it's set up in a way that it can do two things for me. It can either look at an existing job description and come up with recommendations and identify if maybe we're using too much ULA language on how to change it to be a bit more neutral. Or I can just give it a job requirement list and ask it to generate something completely. And it'll actually come up with lines like, It will give examples of different projects we've done that vary in scope or complexity to illustrate a. What it is that we are doing and why it is exciting to do a project with us. And of course, writing job descriptions, you probably know, is super difficult and takes a long time. You're always staring at the screen. Ah, type some more stuff. Ah, and usually you end up making it way too long. And I think the worst sin is when you only make a like list and make a big grocery list like this. With the help of a GT like this, it becomes a lot more story.

TIM: Right.

LUUK: engaging for somebody.

TIM: Excellent, excellent. And yeah, these hiring sessions then help you make that job description, sorry, that job ad as accurate and compelling as possible in the first place. So then your sourcing is better, you're tracking the right candidates, and you're going down the right path to begin with. You'd mentioned it also. In these hiring sessions, part of it is about being able to set up this nice kind of consistent hiring process where, you know, the kinds of questions you're going to ask because they're linked to the requirements for the role and you're not overlapping on questions. You've set it all up with a lot of logic when you've seen hiring processes that lack structure and haven't had that thought-through process. What typically happens in those scenarios? What are some of the downsides or upsides, if you can think of any?

LUUK: Upside, if I can think of anyone. You don't have that, that structure.

TIM: Yeah.

LUUK: So I think there are a couple of pitfalls that really can come from structure. The number one thing would be the lack of repeatable outcome, right? That you get a different outcome depending on the interview that was present. I think also a lack of structure comes from companies not valuing the growth engine, thinking that recruitment is enough. And what I mean by that is that. Sometimes you'll see a lot of companies that, that, that really emphasize it is a very important engine for our growth but then not set up the company in a way that actually enables it. Have an engine you need there. And the result of that is that it's very difficult to schedule interviews, for example, or get people involved and feel like they have a stake in this. And that leads to long delays in actually planning interviews. And I think especially within the field that we are in, in the tech sector, if anybody really raises their hands. With good structure, you can also ensure that your time to hire is shortened and you can schedule those interviews quickly. Maybe automate some scheduling parts of it, right? So when you send me an invite to do this interview, you send, like, a link similar to this, or I could schedule something like that. That's, of course, very easy when it's just one person involved. When there are multiple persons, it gets more difficult, but yeah, when a company has that structure. Keep things at a good speed and round off your entire hiring process within a good amount of time, I would say.

TIM: Yes. And technology, I guess when used right, supports that I, I, I, I sense this is working, running a Luba for six years. I sense sometimes in talent and recruitment circles, there's. A slight mistrust of using technology in some cases, in hiring it, or perception almost, that it could dehumanize the process, like it's automated things away, which it can do. But I also think what could be more dehumanizing than never hearing back? Because the processes are so tediously manual that nobody could possibly get back to every candidate or people slipping through the Those kinds of things.

LUUK: Yeah. I fully agree, but I also think if you utilize automations and AI in the right spots, it actually humanizes the process. So for me, as a recruiter, I hate admin work, like with a passion. I just hate it. And the thing I enjoy most is having conversations and connecting with people and hearing great stories and telling about how great the company I work at is, like what it could mean for them as a professional and as a person. Those things I really enjoy. I don't enjoy it. Tedious stuff. And the more time I'm removed from that, the more time I am able to do things that actually add value for the company that I work at. Which is actually connecting with people, getting them enthusiastic about joining, and explaining opportunities. And pushing people forward into the hiring pipeline and that, with AI, you are more enabled to do something like that. And also it has a great benefit, of course, to the company you work at because, so I've—I'm a very data-driven guy. I just love it. And especially with all the teachings I've had over the years at the previous firm I worked at. So I even create my own models about recruitment and, like, the capacity you need as in recruitment professionals or. Based on historical data and then also on the data that we need for this year, for example. And then, depending on how you time-box your time available as a recruiter team, it calculates how much full-time equipment you would need to actually fill your hiring needs. And the first thing I noticed when I built that model was when I. Situation I had a different company where the ED, the full-time equipment requirement for recruitment, was like seven or eight full-time employees. And by getting some automation done, it got produced to two and a half.

TIM: Wow.

LUUK: And That tells you a lot about how much. Bullshit work you actually have to do when you don't do all the automation stuff because it sounds a bit better. Oh, you need to let go of a lot of people. No, those people weren't even there. The people that were there were just overworked. 'Cause they had to do so much extra and weren't able to achieve their goals because they got stuck in these tasks that weren't generating value.

TIM: Yep. Yes, clever. Automation for the win for sure. You'd mentioned, yes, being a fan of kind of data-driven processes and metrics and whatnot with AI in hiring. Then, where do you see the biggest opportunities in the next kind of couple of years? What would you like to see AI do in hiring, and maybe where could you see humans still doing a better job?

LUUK: So obviously humans doing a better job is in the face-to-face interviews and really making a connection with new employees. I always say for new colleagues, you go this far, but for French, you go even further. And I think that's what humans. You create connections with people, you create the feeling of belonging, and also you. And AI will never do that. It's just a machine. I think what AI is incredible at, especially now when I look at what I appreciate most in the recruitment field when it comes to AI products, is about identifying opportunity. For example, when you have, especially these larger enterprises, they can have hundreds, maybe thousands of roles open. I. Identify coming. Might hiring be somewhere? So there are a lot of different applicant tracking systems right now that are developing systems, and some of them are live, and some are still in beta or something. Where it's coming in will automatically also be matched to other opportunities with your company or if you are opening up a new requisition. It also identifies maybe some key candidates that you've spoken to in the past that might be a match here. I think those would be very beneficial if that's more utilized because, and knowing for effect in the past that people that I've spoken to before a year later could turn into hires, but keeping track of those people manually, again, is super tedious and hard work, getting some AI assistance in that. Super beneficial. I think it's beneficial in both ways, right? You create. An easier way of identifying the opportunities for people that might want to work for you. And it also identifies the low-hanging fruit for you as a recruiter, as a company where you can find those hires.

TIM: Yeah, that's. A great one and should be quite easily measurable as well. It's almost like a new channel. It's like a kind of remarketing channel or rematching channel or something. And so figuring out that it's working and it's valuable will be pretty immediate, I would've thought. Have you seen, you mentioned, like, some TS is trying to roll this out at the moment. Have you, like, played around with any or seen any of those beta tests?

LUUK: Yes. So I know for a fact Team Taylor has it already. And another platform I'm a big fan of is Recruit from Talent. And I know that they are building it right now, and it will go live probably within the next quarter of this year.

TIM: We were talking about AI use cases in hiring, and I think I was asking you about whether or not you think AI is going to be used in the actual core selection and decision-making of who to hire rather than humans or at least partly AI-driven. Woo.

LUUK: Yeah. Because that decision, part of the process for me, is just, I cannot fathom having the trust in a thinking rock to make that decision for you rather than trusting your own knowledge and experiences and also the relevant context around the company that you work at, the position, and the people you've talked to. There are some companies that have experimented with this, right? With some varying results. Mainly getting the results of just hiring more of the same people instead of also having folks or maybe people that add to your culture. Yeah, I hope that will not become mainstream, that AI will make a decision for you. I think so too. Will be more driven to create resumes than the AI.

TIM: Yes, exactly. That will happen. I guess the current problem is, though, I think that maybe we aren't quite as good at making hiring decisions as we might like to be. For example, I'm sure there have probably been similar studies in Holland. There was one. In Australia a few years ago, the University of Sydney got all these job A's and all these resumes and basically just changed the names on the resumes and applied for lots of jobs to then test. Was the name itself a factor in whether or not this person would get a callback? And they tested Chinese-sounding names against Anglo-Saxon-sounding names. To cut a long story short, they found that if you apply to a job in Australia, on average. With a Chinese name, you would have only one-third the chance of a callback as with an Anglo name, which is pretty dreadful. And they did a good job at controlling for all the other factors and variables you might think about, like language and anything like that. And so it's like a pretty devastating result. And I feel like maybe some kind of well-trained, controlled AI could maybe do a better job or more consistent job in that selection bit of the process. What do you reckon?

LUUK: Yeah. But so is that a decision that gets made before they are spoken to? I'm wondering.

TIM: The decision from the AI, you mean?

LUUK: No, but when you're saying that people with, for example, Chinese names have less chance to be hired. Are they being spoken to, or are they applying and then rejected immediately?

TIM: It's specifically from the application to this, the callback rate. So they measured the rate at which the resume would get either an email or callback.

LUUK: So for me, the solution there doesn't necessarily lie in AI. It lies more in a couple of things. I think you know, a lot of people, they are trying to be unbiased. I think being unbiased is impossible. It's human nature to have bias in you. I think it's a lot more reasonable to request people that are participating in interviewing. That they become more aware of those biases and teach them about the different biases there are and acknowledge that you have those ones so you can be a bit more mindful of them and how to maybe prevent some of them. In terms of also helping out in, in, not disregarding canvas. This might be a great fit, but have. So something I did two things to get this done. The first one being that when people apply, they automatically move to a resume check step there. There are about three or four people that are then automatically requested to write an evaluation based on that resume. They cannot see each other's evaluations before putting in their own. So they can't influence each other there. The other thing is that the resumes are fully anonymized, so names are blurred out. Websites are blurred out, and pictures are blurred out. Pronouns are blurred out. Locations are blurred out. You are only basing your evaluation of this person's resume on their resume and nothing else. And only after I've spoken to them, and not then, does that information become visible again. I've been able to have conversations with these people and make an assessment, and whether or not I'm for it, I think those would be the problem more than an AI would.

TIM: I certainly agree with that approach. If you've anonymized and removed all the information or noise that could be used to then result in that bias, then that should be the job done. I agree. That's

LUUK: That's not a manual task, by the way. Yeah, it's, yeah. It's something that's done automatically, so it doesn't require any time for me.

TIM: Yes. And that's implemented in the TS you use, I'm assuming.

LUUK: Yes, luckily it is.

TIM: Excellent. Yes. I think even if you would, you would create some AI that would remove that bias, right? Make the decision for you; it'll always have bias in it. Trained on something that has an opinion attached to it. Yes. We've spoken quite a lot about it. Ai. But one slightly different thing I'd like to ask you about is if you've ever seen any kind of unusual or unconventional methods to find or attract people to roles.

LUUK: Ah, okay. So maybe I can give one of my own that I've done in the past. So I have a lot of different hobbies, and at some point I thought I really enjoyed racing and karting a lot, and it would be nice if I maybe got some more people involved in my hobby. So whenever I spoke to people. Interesting for me, like from a perspective and introductions, of course. At some point I assembled a small list of engineers who also enjoyed carting. So I rented out an outdoor karting circuit.

TIM: Wow.

Luuk and I went carting with 25 engineers. And I just literally had fun with people that were relevant to my work as well. But I. Approach this event as a hiring. But the thing is that it does strengthen your network and your connections. And it did actually end up in three data engineers being hired for me. So it was quite unconventional and even came with zero investment because I didn't approach this as a hiring event. I approached this as just a simple networking, fun activity. So they paid for their activity and had a good time. I gave a round of beers. That's it.

TIM: Not only is that a great sourcing story, it's also a great networking story because I think for a lot of people when they hear about networking, they think about some. Tedious, boring, forced fun drinks event where you're shaking hands with people you don't really want to speak to, but you've managed to turn it into. Yeah. You've got, you've gotten your hobby crossed over with networking. That's a great

LUUK: Yeah. Yeah. I can even say that when I was a little bit younger, I loved skateboarding, and it turns out there were also quite a few IT people at the skate park, so I also made a hire there, strangely enough.

TIM: That's wonderful. My mind's now racing. I'm into football, and so I feel like I should combine football with networking, but I'm also a very crazy player on the field, so I'm not sure that's going to be to my advantage if I go and kick and fight some of the people I'm trying to network with. So who knows?

LUUK: But even in an event like that, that would be great, wouldn't it? Just challenge, like, maybe two different companies and pit the teams against each other, and then that would also create really fun content. So for me, like work needs to be. Enjoyable and fun. And I think I'm at my best when the place I work at doesn't necessarily feel like just an employee, but more of a community help I'm a part of and with great other people. You do, you, you share and work, share a lot of things with and do, and work together. And of course you need to get shit done. But if you're enjoying it. You are a lot more productive. It takes less energy, and I think events like this really add to that perception of it being more of a community rather than just an employer.

TIM: Wonderful. That's such a great example. Luuk, it's been such an easy and enjoyable conversation today. We've covered lots of ground. It's been it's been fun. It's been fun. It's been interesting. Thank you so much for joining us.

LUUK: Yes, it's been fun indeed. Thank you for having me.