Machine Learning Zoomcamp: Free ML Engineering course. Register here!

DataTalks.Club

Articles Slack Events Podcast Books Courses

Recruiting Data Professionals

Season 7, episode 2 of the DataTalks.Club podcast with Alicja Notowska

Did you like this episode? Check other episodes of the podcast, and register for new events.

Transcript

The transcripts are edited for clarity, sometimes with AI. If you notice any incorrect information, let us know.

Alexey: This week, we'll talk about hiring data professionals. We have a special guest today, Alicja. Alicja worked for almost six years at Zalando as a recruiter with a focus on hiring data scientists. I think at Zalando they’re called ‘applied scientists’. At some point, they renamed this, but before it was research scientists, data scientists – you had a bunch of data professionals. I actually met Alicja during a recruitment process as well – she was trying to hire me multiple times. (2:05)

Alexey: Even though Alicja no longer works at Zalando, and I think you’ll maybe tell us a bit more about that later – you don't focus on data professionals specifically anymore. But when I was thinking about who to invite to this interview – who to talk to about hiring data professionals – I couldn't think of a better candidate. Welcome to our event. (2:05)

Alicja: Thanks for having me. It's a pleasure. (3:00)

Alicja’s background

Alexey: Before we go into our main topic of hiring data people, let's start with your background. Can you tell us about your career journey so far? (3:04)

Alicja: I started over 10 years ago in recruitment. Like most people, I kind of fell into it. I started working in a recruitment agency, a very small one, and then I moved onto in-house recruiting. I didn’t become a headhunter, but rather a part of an internal recruitment team at Google, where I spent two years hiring site reliability engineers, which was quite challenging. (3:13)

Alicja: After that, I decided to move to Berlin and I joined Zalando. I actually first joined as a sourcer, then I became a recruiter. The last couple of years, I was also leading a team of recruiters who recruit data professionals and product designers. So that was my journey so far. (3:13)

Alexey: And what do you do now? (3:04)

Alicja: Right now I am part of WeAreKeen, which is an embedded talent agency. So I kind of went to the other side. What we do is go into companies and help them scale their tech teams. We get kind of embedded, so we do the usual things that an internal team would do. We do this in Berlin as well as in Amsterdam, but really, we are operating globally these days with remote work. (3:06)

Alexey: It’s not just data folks, but any tech professionals, (4:24)

Alicja: We are tech-focused at the moment. I'm in the role of a client lead, where it's a bit more on a strategic level – doing project management and helping, advising, and consulting different companies and clients. (4:29)

The hiring process

Alexey: Interesting. We will probably talk more about the work you were doing as Zalando, but I guess the process is similar, where you look not just for data folks, but usual tech professionals as well. So I'm curious, what does the hiring process look like? What are the typical steps in the process? (4:44)

Alicja: A lot of the time, of course, the process overall will depend on the company. But typically, from what I see, especially in tech and for data science, there is usually some sort of recruiter interview as the first step. Not always, but typically. Then there is a technical screening, which is usually one hour-long interview with a data science interviewer and that's much more technical. And then, there is a final round, which is physical on-site interviews, which is why we call them “on-sites”. However, now they're mostly virtual due to the current situation. (5:06)

Alicja: In that final run, typically, it could be anything from two or three, all the way to five or seven interviews. Again, it depends on the company and how big they are. Also, sometimes this process may not be as organized or as structured. After that, obviously, there is the decision round and then the final decision is made. (5:06)

Alexey: I guess even before that, the process also involves sourcing candidates or doing CV screening, right? (6:11)

Alicja: Yes, of course. When I talked about the process, I guess I just kind of explained the recruitment or the interviewing process. But, of course, before we even get into that stage, there is a lot of stuff that has to happen beforehand. First, we naturally need to have candidates who are willing to go to those interviews. Especially when it comes to data professionals, that process could take a while. (6:20)

Alicja: Then, after the interviews, there is another stage where the decision is made, and if the decision is positive (we want to hire someone) we need to make an offer – we need to talk about the conditions and come up with a contract and then onboard that person. (6:20)

Sourcing and recruiting

Alexey: That’s quite a lot of work. And what do you do in this process? What is your role in this process? (7:01)

Alicja: The role all starts whenever there is a need, or a demand – and obviously, the budget also has to be there to hire someone like a data scientist or data engineer. Typically, there’s also a hiring manager, which could be anyone, such as an engineering manager, a data science manager – basically someone in the team who can hire and who wants to hire. This person would contact the recruitment team or myself directly and say, “I want to hire a data scientist, can you help me?” (7:09)

Alicja: That's where it starts. This is where we have an initial meeting with that person to establish what they're really looking for and come up with the job description to make sure that it's aligned with their needs. During this process, we try to establish, “Why do we need this person? Why are we hiring externally?” and “Does this person really need all of the skills that are mentioned?” Sometimes there are certain expectations, but they don’t always match what’s available on the market. The talent market is super tough when it comes to hiring data professionals. It's not easy, there's high competition. (7:09)

Alicja: In a way, it’s kind of a negotiation sometimes with the hiring managers – depending on the hiring manager – as to what realistically is possible and what is not. Going through the job description, making sure it's attractive, and that people will actually want to apply to it. Also, at this stage we decide what the interview process will be, “Who will be the interviewers that will take part? What kind of interviews do we absolutely have to do? Which ones could we maybe omit?” If there is some flexibility in that, of course. So it’s getting into very kind of nitty-gritty practicalities about everything. At which point, the hiring manager will talk to the people who will talk to the candidates. (7:09)

Alexey: That's quite a lot of things. (8:59)

Alicja: That's where it starts. Yes. Of course, then there are two folds. When hiring data professionals, and I think tech people in general, it's a very competitive market, but from my experience more so when it comes to data scientists or data engineers, we do obviously post the positions. Typically, every recruiter would always post the open position, but the number of applicants that apply isn't always aligned, so we can't rely on that because we would probably never hire them. This is where sourcing comes in. (9:02)

Alicja: Although some recruiters do both, some companies have separate sourcing teams where there are people who just source candidates and don't even interview them. But what I did in the beginning of my journey at Zalando was – I started as a sourcer for data science and then I became a recruiter for data science. It goes down to actually actively approaching candidates or people who wouldn't apply because they don't have to. They get messages on LinkedIn, or they may be happy in their jobs and they don't even think about applying. This is why sourcing is often a big part when it comes to hiring data professionals. I'm not sure if I'm answering your question. (9:02)

Alexey: The question was, “What do you do there?” I think you only partly answered it, because I think there's more, right? (10:28)

Alicja: Yes. So, it's quite a long journey. Let's say we post the job and also we typically simultaneously source for the candidates. We start reaching out to them on LinkedIn, but any other ways that a sourcer can find you –you might have experienced this yourself – sometimes it’s even through Facebook, Twitter, Reddit. There's so many – GitHub is another one. Maybe not so much for data professionals, more so for software engineers, but any ways we can find and get in touch with people we think are right, that's what we would do. (10:37)

Alexey: Even on Reddit? (11:13)

Alicja: Yeah. I personally haven't done that, to be honest. But I know some sourcers who are really quite geeky, and they like to come up with new ways. I'm not sure how they do this exactly. (11:15)

Alexey: So if you want to get hired, go to Reddit and post after this interview. [laughs] (11:29)

Alicja: Yeah, I mean. [laughs] Yeah, there are many, many different ways. Also, looking at University alumni, especially for data professionals – if we're looking for people who need to have a certain level, like PhDs, we look at the alumni of universities. We also look at conferences. When I searched for data science people, I looked at that a lot – the machine learning conferences and the papers that were submitted. Sourcing is a big part of the process and it takes a while. It's a long-term strategy of trying to hire data professionals. (11:34)

Alicja: This also comes along with partnering with the hiring manager. At this point, it depends on the sourcer – if there's someone who's just starting as a sourcer or doesn’t yet fully understand what they're looking for – because as recruiters, we are not technical people, we don't have computer science degrees. It also takes us a while to understand what the right profile is. This is why we typically partner with hiring managers, or interviewers, or people in the team, or other data scientists within the company, in order to understand, “Are those the right type of profiles – those who apply and the ones I reach out to?” Then, hopefully, we do have some candidates and we put them through the process. (11:34)

Alicja: I would start with the recruiter interview, spend some time discussing people's motivations, and also discuss the practicalities of what they will be looking for. Then we go through the stages – as a recruiter, I will be the one that takes that candidate through all the stages – starting from the recruiter interview, then technical screening, and then the on-site interviews. As we go, I provide feedback from those interviews, and then all the way up to the offer, including making them an offer. So that's the end-to-end recruitment process, or as we sometimes call it, “360 recruiters”. There are also recruiters who don't do the whole part, but I think more often these days, you just do everything. (11:34)

Alicja: You also then make the offer to the candidate and I think that's something I, as a recruiter, enjoy doing – being able to go through this journey with someone and at the end, make them an offer, and hopefully they accept it, which is even nicer. So making an offer, and then making sure that if they accept it, they also get their contracts. (11:34)

Managing expectations

Alexey: That’s a lot of work. You mentioned quite a few interesting things that I wanted to ask you about. First I wanted to ask about, apart from Reddit which was also quite interesting – but what really got my attention is, you mentioned that when a hiring manager comes to you saying “I want to hire a data scientist or applied scientist.” You first will kind of negotiate with them “What is realistically possible?” Does this happen often when they come with a profile that doesn't exist and you have to manage their expectations? (13:57)

Alicja: [reluctantly] Yes. That's a part of the challenge. This is also why I actually said to you, when you first reached out to me, that I no longer recruit for data science. But still, it's all very close to my heart. In the very beginning of my career as a sourcer for data science, in the first six months I think I was not able to make a hire. For a recruiter this is an important thing to know when you are a candidate, because you need to understand that recruiters are there to make hires, which means they are actually, in a way, on your side –they want you to sign that contract. They will do everything possible to make sure – obviously, within the ethics of the job – but we will do anything possible to make sure you sign it, and that you are happy in your job. (14:32)

Alicja: In the first six months of working as a recruiter, I didn't actually make any hires. I wasn't able to find anyone, and this was quite a challenge. Sometimes a big part of that is the expectations that some hiring managers have. Later on down the line, I was very lucky to have very good hiring managers at Zalando, who really understood the market. Over time, we were able to build these relationships with each other. There was an understanding that if a hiring manager comes to me saying, “I need three principal data scientists.” [laughs] I would say “That's very funny.” But I was comfortable to say that, of course, because I knew that they understand that this is not going to happen in the next month. I would typically say, “Yeah, it's going to be a minimum of six to nine months before we hire the first one.” This is very often the thing. (14:32)

Alicja: I think sometimes it's because hiring managers don’t necessarily have the data science background themselves. It could be that they're an engineering manager, so they have a software engineering background, but they need to hire data scientists for their teams. This is where I, as a recruiter, would need to provide them with more advice and give them insights of the market, “What are the profiles that are available? What is realistic?” Oftentimes, they would understand that and that would be fine. So we would agree on “Okay, what are the must-haves and what can we compromise on?” (14:32)

Alexey: So you have this information. You can say, for example, “Okay, finding this kind of experience with Kubernetes is very difficult in the available data science profiles.” Right? Then if the hiring manager really insists on having this profile, you will need to say that, “Okay, this severely reduces the number of candidates that you would have for this position.” (16:50)

Alicja: With data, obviously, you have to. I think, especially in tech recruitment, but especially when dealing with data scientists, I've noticed this is the thing. You just need to have the data to show them and that's the easiest way to convince someone. Then it's not about my gut feeling then. As a sourcer, and also as a recruiter, you're able to gather this data – LinkedIn also has tools like Market Insights. A lot of recruiters have access to that. Some of us don't, and then you have to get a bit more creative. But you can still show them, at least roughly, “Okay, this is what the market looks like, so the more must-haves you add, the narrower the funnel is, and the less talent is available.” That helps to drive the conversation. (17:18)

Making the job description attractive

Alexey: Interesting. Another thing you mentioned is – you work with the hiring manager to make sure that the position (the job description) looks attractive. How do you do this? By carefully picking the words or how exactly does that work? (18:11)

Alicja: For me, it was always the case where I wouldn't write the job description myself. Because I don't know what specific AI tools or methods in machine learning have to be in the description or how to word it so that it actually makes sense to someone reading it (the data scientist). I would always first ask the hiring manager to draft a job description and we would have some format that has to be followed. Then we would just check that the wording makes sense. (18:28)

Alicja: But I think it's also about understanding that it’s not just about putting the right buzzwords in, but rather being able to really show what the person will be doing. When I was going through sourcing and working with hiring managers, I found out that data scientists, or data professionals, like to know what kinds of problems they will be solving. They get excited about the problems. So being able to write a job description that focuses on “This is actually the problem we have right now and you'll be part of the team who is trying to solve it.” That's what gets people’s attention and excitement and not just saying, “Oh, yeah, we have perks and benefits like free lunches or coffee.” Yes, that's great, but it's not as exciting. (18:28)

Alexey: Free lunches are good as well. [laughs] (19:53)

Alicja: I agree. Yeah. [laughs] But when it comes to making a decision to switch jobs… (19:55)

Alexey: That's not enough, right? (20:02)

Alicja: It doesn't flip the scales. Another part of actually writing job descriptions – this is something that is maybe more recent – there are actually AI-driven tools that allow us to go through the job description and tell us things like, “What kind of audience will this job description attract?” Right now, as you might be aware, there is a big conversation about having more inclusion and diversity, in tech especially. When it comes to gender diversity, for instance, there are certain tools that are able to tell us like, “Okay, this job description that you've just written is really speaking louder to men (or people who identify as men) more so than a woman. So if you're actually trying to balance out your team a little bit, maybe you should change these keywords.” And it gives you an indication of what kind of words you can swap. That's actually something that is AI- or ML-driven, which is quite interesting and it can be helpful as well. (20:04)

Alexey: Yeah, thanks. I think I also used this tool. It would say, “Okay, this phrase (I don't remember the exact examples), will appeal more to male candidates. If you want to be more inclusive, consider changing it.” And then it gives a couple of suggestions. (21:11)

Alicja: Yeah, exactly. It makes it more neutral. (21:29)

Selecting profiles during sourcing

Alexey: Let’s talk more about sourcing – you now posted it, you negotiated the requirements with the hiring manager, you’ve made an attractive job description. Now you go live with this description and people start applying. But you also mentioned that it's not enough. You also need to actively reach out to people. You go to Twitter, Reddit, GitHub, or LinkedIn, and people say, “Hey, you have this amazing job description. How about having a chat?” In this case, how do you select profiles? What piques your attention here? (21:32)

Alicja: Typically, to be honest, there are many tools you can go through. But I think LinkedIn is a good starting point, usually, because you can at least get an idea of who's available. It also has quite interesting filters. I can filter out for years of experience, or the degree levels, and things like that. Then I can put Boolean strings of keywords. It's quite good in that respect. So that will be a good starting point. (22:13)

Alicja: What I would do is look at profiles that matched the initial keywords, or the job description that I have in front of me, or based on the discussion that I had with the hiring manager, the most. First, I try to narrow it down to whatever is the “ideal” case scenario. This way I know where those people are and who they are. Then I verify that with the hiring manager to see, “Okay, these are the initial results. I haven't reached out to those people yet, but what do you think? Is this profile correct?” Because, as I said, I might not really know if I'm just starting out in the field. (22:13)

Alicja: So what stands out? Something I look at quite a lot when going through LinkedIn and sourcing is the experience and also the education. I think this is something that has to be balanced when looking for data science people or data professionals. Often a degree like a PhD, or at least a Master’s, is a must. Not always, but that would be something I would look at. As opposed to if I was looking for software engineers, for instance, where that necessarily wouldn't be the case. But when it comes to data science, it would often be a requirement. I tend to look at what kind of degree it is, “Is it computer science? Is it physics? Is it biology?” Whatever it is, I then check if this matches the position. (22:13)

Alicja: But there must also be a balance with the amount of experience and what that person did. Not just the years of experience, but what they write in their responsibilities. I think the worst scenario for a sourcer is coming across a “skeleton” LinkedIn profile, where there is just the company they worked for and their position title – but nothing in the description of the tasks or responsibilities that they do. Sometimes people just put the description of what the company does. I know what this company does [laughs] I already did research on them, thank you very much. I wish to know what you are actually doing there. (22:13)

Alicja: Of course, I think some people do it on purpose because maybe they don't want to be contacted by recruiters. I understand that. Then I have to judge if this is something I still want to do or not. But then that also means my message to that person is not going to be very personal because I really don't know if I need this person or not. It's more like a stab in the dark. So it’s looking at exactly what their responsibilities are, what this person did themselves, not just what their team was focusing on, but what they did themselves. Did they build those machine learning algorithms and models from scratch or not? Then I combine that with looking at the degree and, obviously, the keywords. As a recruiter, those are the things I can rely on. I can't read a paper that someone wrote for a PhD and understand it. Often, I don't. Sometimes I do – maybe the summary at the beginning, but very vaguely. There are very few things and not always getting it right. (22:13)

Profile keywords

Alexey: Do you remember what kind of keywords you put in there? Let's say if you were to look for a data scientist right now, what kind of keywords would you put in the search tool? (25:56)

Alicja: “Machine learning” or “AI” or “ML” or “deep learning” or a few other a bit more popular ones like “algorithm”, but that also depends on the job description. For instance if you're looking for a computer vision specialist with deep learning skills, then yes, I would put “algorithm” but maybe I wouldn't put some other mathematical models and things in there. (26:07)

Alexey: So you take them from the job description, right? (26:47)

Alicja: Exactly. Yeah. That would be my starting point – the must-haves. There’s like three or four must-haves and I would just put it in the string. Sometimes it would give a very broad pool, but then I would narrow it down with the degrees and the level of experience. (26:50)

The importance of a Master’s vs a Bachelor’s degree vs a PhD

Alexey: How important do you think it is to have a Master's degree? Let's say somebody only has a Bachelor’s degree level of education. Or maybe they don't have any degree at all. How important is it? (27:10)

Alicja: I don’t really think there’s that much difference between a Bachelor's and a Master's. I feel like there was always a bit more distinction if it was a requirement coming from the hiring managers, “Do they need to have a PhD or not?” If it’s not, then it wasn't specifically “Oh, that means they must have a Master’s.” It was more like “It could be a Bachelor's or a Master's.” But I think there is more of a distinction between Master’s and PhD – sometimes that is important. (27:24)

Alicja: For instance, it would be a must for a team that is a very research-heavy team, where maybe they weren't working on a specific product. For instance, there was a research lab within Zalando, but they didn't work on a specific product and delivering that. But they needed to have a PhD and papers – that was the point of the team, more or less. Really high level from a recruiter perspective and I don't want to undermine that. But then for other teams maybe they didn't really need people who have PhDs or even a Master’s. Of course we still want them to have a solid education in terms of understanding the maths behind some of the tools they would be using and the algorithms. But they don't have to have Master’s. That was often also something we discussed. (27:24)

Improving CV

Alexey: Okay. I also prepared a couple of questions about CV screening and I think we partly covered that. I assume the process for you of looking at a LinkedIn profile is similar to looking at a CV that you get when somebody applies. Is it similar? You mentioned for a LinkedIn profile, you look for exact responsibilities – what exactly the person did. Right? I think this is still important for a CV. But what else is important for a CV? How do you look at a CV of an applicant to decide if you want to continue talking with this candidate or not? (28:41)

Alicja: I think a CV is only slightly different in the sense that on LinkedIn, there is a certain format. Typically, you will see the experience, the role, and the responsibilities first. After the name and the picture, of course. Whereas in the CV, sometimes I noticed that people would maybe put their education first. Working for an in-house recruiter for a company that needed typically mostly senior people (in terms of their experience) the experience would be the first thing I would look at and at the education after that. Of course it depends. If you are a recent graduate who doesn't have much experience, maybe it is better to put the education to highlight that as a first point. (29:18)

Alicja: But I think experience typically beats most of that. Even if it's just a six month internship, I would already put that on top of the CV, because that's what I would look at. But CVs and LinkedIn are very similar. It's looking at experience and education. Again, with the CVs, I would also often see that there will be a title and there will be a list of responsibilities, but they will be very vague. Or sometimes people put just a list of tools. But… it’s just a list of tools and I’m not really sure what they did with them. How often did they work with them? I have no clue. (29:18)

Alicja: Sometimes I also see in CVs – maybe not so much on LinkedIn, but that could also be the case – that people put a lot of those buzzwords, which can make things very confusing. It used to be very confusing for me, because then I was like, “Okay, well, this person mentions all of the things that I have in a job description. It looks fine.” But then it would turn out that they actually wouldn't be able to pass the first interview. I was like, “Okay. Did I do something wrong? Maybe I should have explained better.” But I was also lucky enough to have built a sort of network of data scientists within Zalando, who were very, very helpful. (29:18)

Alicja: I remember specifically Nikhil Brown, who I think was on your podcast some time ago. He would sit with me in my very first year as a sourcer. Every week, I think for an hour, or maybe even more, we just screened CVs because I had no clue what I was doing. I was like, “Look, soon it's gonna be six months soon, and I'm not sure if I'm going to pass my probation here because I haven't made any hires.” So that was very helpful as well for me – to see how he looks at the CVs from the perspective of someone who actually does the job. That was very helpful, because I learned to read between the lines, or rather learned to read beyond someone just putting in the buzzwords. Sometimes people would put the buzzwords, but there would be nothing else. Or it could be the opposite – they wouldn't actually showcase or highlight what they did and then they could miss out. Maybe you will get rejected just because you didn't put something in your CV. (29:18)

Alexey: What can people do to make their CVs more attractive? The first thing you said, that people should move their education down and put in very nice work experience. What else can candidates do to be more successful at this stage? (32:23)

Alicja: They should be very clear about their responsibilities and what they did in their current job, in their previous jobs, as opposed to what the team did. Of course, I don't mean “don't give credit to the team” and don't own up to things that you didn't do on your own. But do highlight “What was your part in that team effort?” At the end of the day, that's something that we need to know. That really helps a lot. It also helps a lot with driving interviews and for the interviewers further down the line, because then the interviewers can ask you more informed questions and avoid asking you questions that are totally irrelevant, which you could also get confused by. (32:40)

Alicja: So be very specific about your tasks and responsibilities – what you accomplished and the things you did. If those things were part of a team effort, of course, also mention that. But this is still important to add. More so than the title, I think. Because titles can be so confusing. This was also a big part of the confusion when I started as a data science sourcer six years ago, the title could mean anything. A “data scientist” could be anyone. (32:40)

Alexey: Yeah. Not many things have changed since then. (33:50)

Alicja: I was hoping that it did. I thought it would get a little bit better. [laughs] (33:54)

Alexey: A little bit. But still, you really have to look into the responsibilities to understand “Oh, okay. This is the kind of data scientist you mean.” (33:58)

Alicja: Just to answer your question shortly, because I don't want to go over time, but another thing about the responsibilities being very clear – put the month and the year, not just year to year, because that could be up to a 12-month difference sometimes. Also practical things, like making sure there are no typos, but I think that's kind of quite obvious. (34:07)

Alicja: If I could offer my opinion, or perhaps my “preference” – I have a lot of recruiting colleagues who also mention that there is this “Europass” format of CVs… it’s horrible. Very, very difficult to read. I have no idea where anything is and I think it’s related to the EU somehow. Not sure but… it's something I would avoid. One thing I want to also mention to avoid, because I know in Germany in particular, it's kind of more of a tradition to put your photo on the CV. In some countries, not just Germany, I've seen this, that you add your photo. I know that in some more traditional companies, this is required. So you actually have to add your photo. But the trends I see now and what I think is more beneficial to people applying – is to not attach the photo. (34:07)

Alicja: Also don't put your date of birth and things like that. We are all biased, and it's unconscious bias. Not intentionally, but when you see a photo first, you immediately start forming assumptions in your brain without knowing it. This also helps you as a candidate to reduce that bias, and hopefully be more successful. Of course, you have to judge for yourself. As I said, some companies will require that and that's fine. But I think for a lot of maybe smaller companies, startups, or tech companies, I would avoid adding your photo. (34:07)

Interview with the recruiter

Alexey: Yeah, thanks. So. Now you look at the CV, you decide to move forward with a candidate. The next step is an interview with a recruiter – with you. Right? So what does this look like? What do you talk about there? (36:08)

Alicja: I think my interviews are quite short. I bet that every recruiter does it differently. But typically, recruiters are not going to ask technical questions. We can't. At Zalando, we did try at some point in the very beginning, to come up with very short screening questions on the very basics of machine learning. I could ask them and verify the answer – a kind of ABC answer. However, everyone was passing them, so I decided that it doesn't make any sense to keep doing it. (36:25)

Alicja: My focus would be on the so-called soft skill, which is more about “What has the person been doing?” Sometimes there are some things in the CV, like gaps, that I would also like to clarify. Or if the CV is a bit more generic, then I would ask “What were your responsibilities and accomplishments in your current or previous jobs?” If the PhD is important, “What was it about?” And I would sometimes ask, “Can you explain it to me as someone who is not technical?” For data scientists, maybe not so much for less experienced ones, but for the more senior roles, it's also a part of it. It's not just about the technical ability and skills, it's also about being able to explain those complex concepts to people who have no idea about machine learning – to be able to convince them sometimes that “This is not a realistic solution that you have in mind. We can't do that.” I would test this on myself, like, “Can you explain to me what you did in your PhD in words that I can understand?” I really liked that because I also learned quite a lot. I learned to actually really like data science and talking to people who did it. It's really fascinating. (36:25)

Alicja: I would typically ask behavioral questions as well. So things like “Can give me an example of a situation when you had to work with a difficult stakeholder? Why was it difficult?” This, I think, could be a difficult part because I would see a lot of people answering in very general or hypothetical scenarios. So try to answer by giving me an actual example from the past and walk me through your actions and what you did in that situation specifically? What was the outcome? So, these are called behavioral questions, because we want to check for behaviors – we want to see if the person will fit with the culture of the company and the values that the company has. So that's why we ask them. (36:25)

Alicja: Then there are practical questions. At the end, I typically ask about the notice period – if you're currently employed. How soon you could start if you decided to join us and then, obviously, salary – what are your expectations? Also how active are you interviewing currently? Because I might be sourcing people who are passive, but so are 15 other recruiters and sourcers from other companies. Sometimes it can end up that I speak to them in the very beginning of the process, but at the end, they already have other offers on the table. Therefore, I want to verify in the very beginning “How advanced you are in those stages?” Because there is still a process to happen, so “How much time do we have left?” Those would be the practical questions. (36:25)

Salary expectations

Alexey: When it comes to salary expectation questions, I heard this advice – “Never say the number first.” So when somebody asks you, let's say I'm a data scientist and you're a recruiter. We are having this initial screening. You ask me, “What are your salary expectations?” And if I follow this advice, then I would say, “No, I will not tell you. Tell me – what is the salary range?” Then the recruiter would say, “But we want to base…” and this Ping Pong starts. So what do you think about this advice? Is it good advice? Should people follow this? (39:53)

Alicja: It depends. But what I would do before answering first, is actually say, “It can be super helpful to know the expectations.” However, I don't think that it’s necessary for you to say what your current salary is. This is something I follow myself. When I was interviewing recently for a new job, I made the point of not answering that question. And it's absolutely fine not to say what your current salary is. It’s no one's business. Also, in most European countries, at least, your future employer won’t be able to check. Even if you say some number and it's not really that – they won’t be able to check that. And they shouldn't be able to – it's personal data. So ignore that question or just say, “Look, that’s not something that I'm willing to share.” Sometimes recruiters are told that this is the question they have to ask and they will ask you, but just know that you don't have to answer that. (40:33)

Alicja: When it comes to the expectations, though, that's a bit trickier. We want to know if your expectations are totally off, and I mean like 20,000 off in terms of the annual number. If this is so, then we probably shouldn't be wasting each other’s time. As usual, it depends on the company. I speak to candidates every day, so I know there are some companies – how their structure internally is based – that if you interview them, it depends on how good you are at negotiating. So if you're not good at negotiating, you will get whatever you asked for in the beginning and that's it. Someone told me, “Yeah, they gave me this salary and I learned later that my colleague in the same level earns this amount more.” When I told this to them, I was like, “Well, yeah. You should have negotiated better.” Some companies still do that, I think. Typically, smaller companies. (40:33)

Alicja: That's why it's good to actually ask, “How does your salary structure work? How does the leveling and structure work in this company?” before you answer the question of salary expectations. If it is the type of company where it all depends on how well you interview and what you ask for, then don't tell them your expectations. I'll be honest – don't. But in a lot of companies, like the bigger ones, there is a level system in place. That means that to every level, whether it's Junior, Mid level, Senior and above – there is a band attached. I actually say this to candidates right away – no matter what you're going to say, even if the number is very low, I will tell you what our band is. But, just so you know – just because you asked for a lower amount, we're not going to give you that amount. We will still put you on that range if we feel like you are truly a senior or mid level. It will be very unfair if we give you a lower salary and I think a lot of companies try to avoid that because it is unfair. It leads to a weird atmosphere between the members – not a good thing later on down the line and in the long term. If it's a company that has established ranges and leveling, then I think it's okay for you to share your expectations. You don't have to actually say a specific number – you can give them a range. That's already sometimes good enough. (40:33)

Alicja: Also, what you can say is that “This is my initial estimate of what I want, but this may change because I'm interviewing with two other companies. I will also do my research.” This is sometimes what people also tell me – “Okay, well. I thought about this number.” But two weeks later, they maybe got an offer that's higher than this and now it's changed. And that's also fine. This is life, right? As a recruiter, we expect that. This is not set in stone – I'm not going to hold you to this at the end. So like a month later, turn around and say “Oh, but you said this.” This would be wrong to do. In the data science recruitment market, it would be weird for me to expect that people will not have other offers. Of course, it depends. Sometimes, if you're just a graduate, it's probably harder to find a job. But people change their minds and expectations change and that's fine. As long as you're open about this from the beginning. Say that in the beginning, “This may change.” And keep the recruiter posted. That's all we want. Just let us know so that we can counteract and try to still maybe do something about it. (40:33)

Alexey: You mentioned, if I would say, “Well, I'm having this conversation with the recruiter.” And then I give a number that is quite high and you see, “Okay, there is no way we can match this with what we have.” But maybe the reason I told this high number is because I have no idea how much to ask. I just picked a random big number and then said it. What are the dangers here? One danger is that the company might decide not to continue, right? (45:05)

Alicja: Yeah. Well, if someone did that – actually gave me a number but it was very high – I would ask them “Okay. Can I ask you what this number is based on?” Because sometimes people, as you said, have no clue. They just throw a number out. But sometimes people look on Glassdoor, which allows you to see salaries of people from the same company you're interviewing with. There are a few other tools as well. Some candidates recently told me where you can check salaries globally, for many different companies. From there, you can already gather some sort of a range. (45:37)

Alicja: Sometimes people base it on, essentially, the fact that they see salaries being posted on Glassdoor or other sources and that's a bit more informed. But if it's just a stab in the dark, and it's very high, I would ask them “Oh, is it based on other offers that you already received before? Or is it based on just anything?” I will still try to somehow figure out if we can move forward. As I said, as a recruiter, I really want to hire people. But if someone says like, “No, no, no, this is a must and I'm not moving from it.” Then the danger is that we would just have to say, “Okay, then I guess it doesn't make sense to move forward with the process.” But I have to say, if it's a very, very big gap. (45:37)

Alicja: To be honest, these days I don't very often come across people having very wide expectations, where there’s a very, very, big gap. Usually, it could be maybe a couple of thousand in the annual amount. Sometimes it could be outside of the range for that level, but that could mean that “Okay, maybe we can discuss with the hiring manager and find another way to compensate for that.” There are those ways and I think every recruiter would try it, if it's possible. But if they tell you that it's not, then it's probably true. (45:37)

Advice for “career changers”

Alexey: Okay, thank you. I see that we have quite a few questions. I think it's time we try to cover them. The first question, from Amen, is “How can career changers get in the eyes of a recruiter? What would be your advice to people who are changing their careers?” Let's say they want to get to Zalando or some other company you're hiring for? What advice would you give them? (47:36)

Alicja: Yeah, I think this is quite tough. I know that a lot of companies want people to have a specific degree, even if it's just a Bachelor's, but from a university and in computer science. And then if you come from a different background, and you’re changing careers from something not even tech related, it's really challenging. A lot of people do that, like, do some of the courses on Coursera and then try to swap. But I think it is very difficult, because experience is what matters the most. And unfortunately, that's difficult. (48:03)

Alicja: I would say try to gather experience. Even if it's an internship, an apprenticeship – unpaid. This is, again, something that I hope will change, but for many people, it's not easy to just do an internship that is unpaid for six months because you have bills to pay. That's a difficulty that I totally understand. Another advice I would give though – try to network as much as possible with people who are data scientists in those companies. Whether it's meetups or any kind of communities you can find. But also connect with recruiters on LinkedIn, because there are a lot of recruiters who are in-house recruiters and maybe they work for those companies, but the companies are not really willing to look at people who are career changers. (48:03)

Alicja: Then there are also headhunters from recruitment agencies, who may typically have worked with many different clients. So they will work with many different companies and they have a wider net and maybe there are some opportunities somewhere. So the more recruiters you also connect with, the better. Then, of course, make sure your LinkedIn profile is as clear as possible and you list everything – any kind of practical experience you had – make sure it’s there and it’s explained clearly what you did during that. But yeah. It's a tough thing, especially when it comes to data science hiring, I've seen a lot of the hiring managers saying that “We need people with five years of experience at least and it has to be from a similar company – working on the product, in the product team.” It's sometimes not realistic. (48:03)

Cover letters

Alexey: Yeah, thanks. What do you think about the importance of a cover letter? This is a question from Alma, “How important do you think it is to have a cover letter in Berlin?” Or should it be more based on CV, where the CV is more important than a cover letter? (50:33)

Alicja: If the cover letter is obligatory during the application process, and sometimes it will be, then, of course, it's important. That means that the recruiter or hiring manager, are both, will read it. I would say a lot of the time, I wouldn't read them because I just didn't have time. I think a lot of recruiters are in this position, to be honest. But some of my hiring managers did. So sometimes they will mention, “Oh, this person mentioned that in the cover letter.” But I would say that I don't find it very important unless it's an entry-level position. (50:53)

Alicja: If you're a graduate and there is a junior level data scientist position, then typically a cover letter would be something that the recruitment team would look at – because you probably don't have that much experience to put in your CV, so they want to understand your motivations a bit more. But if it's more seasoned or a more senior-experience level position – unless it's necessary, maybe there is a star that says this is mandatory – I wouldn't put it. (50:53)

Alexey: Yeah, I must confess, as a hiring manager, (I often take part in the hiring process) I also don't look at cover letters, to be honest. I assume maybe the recruiter took a look. Sometimes I also try to avoid looking at the CV, not to bias myself. So before the interview, I just talk to the candidate. Then after the interview, I look at the CV. I assume that the CV screening already happened by the recruiter or by somebody, so I try not to look at the CV as well. (52:01)

Alicja: Yeah, that's a good one. It’s like I said earlier about the photo, for instance. But there are so many other things that people put on the CVs and it can really sometimes bias you without you even knowing. If this person was already screened by a recruiter and you trust them that this is probably the right candidate to talk to, then yeah. I also think it's an interesting idea. Yeah. (52:36)

Alexey: Yeah, but probably for a recruiter, it's not the best way of dealing with CV screening. [laughs] Right? (53:03)

Alicja: Unless we need your help. So if I would still be on the fence about some profiles still, then I would ask the hiring manager. I would ask, “Yeah, can you please look at the CV?” But I think if I'm kind of quite confident and calibrated as a recruiter already, and I kind of just know, “Okay, this candidate looks good” then I think it would be fine that you don’t look. Again, it’s a way to try to reduce bias a little bit, which I think is good. (53:11)

Data analysts

Alexey: Do you also need to hire data analyst profiles? (53:40)

Alicja: I haven't myself, but at Zalando, at least – at the time when I was there – it was a separate job family. But I do see that this is also becoming quite challenging. I hear a lot of my colleagues and the recruiters I work with who mention that it's sometimes really a struggle to find them as well. (53:45)

Alexey: There is a question from Amin, “Do you have any tips for those who are looking for data analyst jobs?” (54:09)

Alicja: No, I don't think so. It's kind of like with the data science title that we discussed – sometimes data analysts can also be a little bit confusing from one company to another. It could be even that there is no data scientist title, but actually data analysts do what data scientists would do in other companies. So I think it would really depend on the job description that you see. But I think I probably don't have that much information about that, I'm afraid. (54:20)

Alexey: The recruiting advice is probably very similar – except the screening phase. Maybe you would look for different kinds of keywords or maybe different kinds of responsibilities. But the rest of the process, I assume, would follow similar steps, at least from the point of view of the recruiter. Right? (54:51)

Alicja: Exactly. Yeah. (55:02)

Double Bachelor’s degrees

Alexey: What about somebody having two Bachelor’s degrees? For example, one in IT and one in account management. Do you think that's a good thing? Would it make candidates stand out or not? (55:05)

Alicja: I think that would be interesting, but I'm not sure if it'll be something that would weigh the scales. I think it's important that there is a technical degree, whether it's computer science-related or related to machine learning. That is probably something I would look out for more. But if there is also another Bachelor's degree in non technical field at the same time, that's usually something interesting. I would look at it, but it wouldn't weigh the scales in one way or the other, I think. (55:27)

Alexey: So it would be “Okay, interesting.” But no more than that. Right? (56:01)

Alicja: Yeah. I mean, sometimes it could be a good thing. Again, it depends on the role, to be honest. It really is a bunch of assumptions, right? If someone has account management experience, I'm making assumptions like, “Oh, maybe they're good at presenting and have soft skills like managing stakeholders,” But then that's just an assumption. I'm not sure. I haven't been particularly paying attention to that, I have to be honest. (56:04)

The most difficult part of hiring

Alexey: Okay. There is a question from Rado, “What is the most difficult part of hiring a data professional?” (56:35)

Alicja: For me, in the very beginning, it was the ambiguity about the titles. I heard this from everyone. The biggest difficulty was that there was suddenly this “data science” title that just exploded. It just became a thing and everyone was putting that on CVs, or at least it seemed to me that way. On their profiles, everyone wanted to do AI or wanted to do deep learning – they would just put it everywhere. Then the companies would also want to say that they are AI-driven and they use data, deep learning, and machine learning. So they would hire people and give them a data scientist title. Then I would hear from people when I would interview them, “I was given this title, but I am really just a business analyst. I have nothing to do with my degree, which is on deep learning.” That's why they would want to move. (56:45)

Alicja: So there was a lot of confusion and my difficulty was like, “How do I know if someone really worked as a data scientist in terms of what the current company I work for thinks about that?” And then also having to explain to them, especially source candidates. I remember we would go to a neural information processing conference with Zalando and a lot of people would come to our booth. It was me and my colleague from branding, like non-data scientist people and it was like, “Why are you here? What does Zalando do with AI?” So people would be very suspicious. Also the candidates, because they had those bad experiences with companies telling them “Yes, you will be doing deep learning.” And then they didn't. (56:45)

Alicja: I had to really become very good at explaining “Yes, this is how we use it and these are the problems you will actually be solving.” So this disconnect between the title and what it really means – that was the biggest hurdle. The second one was that there is a lot of competition. The right type of people that we wanted to hire would have many, many offers. Weekly, they will get bombarded with messages and it was about how to convince them to join that particular company. It was so competitive. Typically they always have two or three other offers on the table with all sorts of different pairings. So that's the second one, I think. It's quite scarce talent, I would say. (56:45)

Alexey: Do you have time for a couple of more questions? (59:19)

Alicja: Yep. (59:22)

Alexey: Yeah, I should have started with questions earlier. I realized there is a pile of them. My apologies for not starting earlier. (59:24)

Alicja: No worries. Happy to answer. (59:27)

Coursera courses on the CV

Alexey: Do recruiters consider portfolio projects, like from courses from Coursera? When you look at CVS and you look at projects, do you think, “Okay, this person has a project from a Coursera course.” Do you think it's a good thing or not? (59:30)

Alicja: Depends on the hiring manager. For some of them, it would be like, “It's nice,” but again, like with the double Bachelor's – it will still be about the experience that they had. Specifically with the specific models or tools and how they implemented them. So the experience itself is more important. I have seen a lot of people doing the Coursera course of Andrew Ng, I think. I have seen a lot of people do it and it’s in a lot of the CVs. I think maybe if it adds value for the person and it helps them, even if it is to actually do the technical interviews sometimes, and refresh your knowledge then that's great. But it's more like a “Nice to have.” (59:48)

Alexey: Okay. That will certainly be helpful for the person to pass the interview. (1:00:41)

Alicja: It might be. (1:00:45)

Alexey: It is very likely – at least as I remember, the questions I got during the interview for Zalando – some of these questions were covered in that course. I also took that course and that course was quite useful to me in the interview. But as a thing on your CV? Maybe it's not the most important thing that recruiters care about, right? (1:00:45)

Alicja: No, unless it's really specifically said that someone has to have it. It's just never happened to me, so I don't want to speak for everyone or every company. I think there are some companies that are more open to that, which is great. Then definitely put it in. And I think in general – put it on your CV. It's something you accomplished and something you did from beginning to end. That also says something. You committed to doing that and putting in those hours. That's in and of itself, I think, is also something that adds value. (1:01:07)

Making a good impression on recruiters

Alexey: Okay, thanks. Let's take the last one. The question from Rondo is, “What advice would you give to data professionals to make a good impression on recruiters?” (1:01:37)

Alicja: It's not that difficult. [laughs] No, just kidding. I don't know, I particularly enjoyed recruiting and interviewing data professionals after I got to know them. Like for Zalando, I managed to get to know a lot of people and understand things that they are excited about and things that excite them. So if you want to impress a recruiter, I think try to answer the questions as much as possible. Try to think about the fact that you're speaking to someone who doesn't understand anything, probably. I don't want to talk badly about recruiters – a lot of us do try to understand machine learning, and we can sometimes tell the difference. But just try to explain as much as possible in words that are maybe not as technical. (1:01:49)

Alicja: Sometimes I would speak to someone and I would ask a question that would be more open-ended, but they would answer with ‘yes’ or ‘no’ and that's kind of, “Okay, I really need to know a little bit more to understand your motivation.” Or like, “Why were you doing this?” or “What did you even do?” So coming prepared for the interview – and that doesn't necessarily mean knowing everything about the company you're interviewing at – but be prepared and have some specific examples as I said about the behavior questions. They are typically not very varied – they are usually about similar values. Companies want to know if someone is a team player, so the questions will be around that. Or if you were more in a senior role and leadership, “How did you lead the team? How big was it? Can you give me an example of how you would grow someone?” So try to be ahead of that and try to think about it. (1:01:49)

Alicja: The interviewer will probably ask you questions that need specific examples, so try to come up with some specific examples as opposed to hypothetical scenarios, like, “What would I do if I was in that situation?” I think that's kind of like wishful thinking. Also, come up with questions. If you're motivated and you’re actually interested in joining the company that recruiters are representing, then ask questions about the company or the culture. If someone's not really asking those questions, I would sometimes wonder, “Are they really interested in this company? Maybe they’re just interviewing with two others and that's what they're waiting for. So this is kind of ‘why not?’” Another thing, I think it's also about commitment. Sometimes, if things change, just explain. Of course, as I said about the salary expectations – things change. It's about respecting people's time, I guess. I know recruiters can also be quite difficult about that – or not very good with that. (1:01:49)

Alicja: I understand that there are some recruiters who don't really keep you posted. They don't say what happened. You never hear from them after you were rejected. But if you’re interviewing with many companies and some company makes you an offer, let the recruiter know as soon as possible “Now I actually don't want to interview with you anymore, because I got this offer.” That's fine. At least you're not wasting their time. Be open about that, “Yes, I am interviewing.” We are expecting that you will be interviewing with other companies. I'm not gonna be offended by that, like “Oh, no, this person is also interviewing there. That's bad.” (1:01:49)

Alicja: Then, with the offers, I think – once you accept an offer verbally with a given company, of course, it’s legally binding, (even though if we make an offer to you, as in most of European countries, it actually is legally binding) but I would expect then that you will honor that and that's your commitment. But I did have sometimes instances when people would accept the offer or even sign the contract, and then a week later say “Oh, I got this other offer now. Can we negotiate this whole thing again because I got more money? (or something else happened). That makes me wonder if the person really wants to be here. It's totally fine if you get more money from another company, and you want to join them – just don't accept the offer before that. So there’s different situations, but at least communicate as much as possible and explain. That's what I would do. (1:01:49)

Wrapping up

Alexey: Yeah, thanks. Thanks for the advice. How can people find you? (1:06:44)

Alicja: LinkedIn. (1:06:47)

Alexey: LinkedIn, ok. So just put your name and you will be there. Okay, so thanks a lot. Thanks for joining us today. Thanks for sharing these stories – these tips. I also apologize that we didn't cover all the questions. So if you want your question to be answered, you can go to our Slack and ask these questions in our #careers channel. You can do that and there are a lot of people who can also answer them. Thanks a lot, again, for joining us and thanks, everyone, for joining us today and asking questions. Thanks for taking part in the discussion. I wish everyone a great weekend. (1:06:49)

Alicja: Thanks for having me. (1:07:00)

Subscribe to our weekly newsletter and join our Slack.
We'll keep you informed about our events, articles, courses, and everything else happening in the Club.

DataTalks.Club. Hosted on GitHub Pages. We use cookies.