What's wrong with hiring in data and tech, and how do we fix it?

Transcript

Hello everyone, welcome back to the second episode of my video blog. Today we're gonna talk about hiring, particularly about hiring for the data team. As you might remember, my name is Ilya, and in this blog I talk about data and leadership while smoking pipes.

Today I want to talk about hiring, everything from sourcing the right candidates up to onboarding in your team. It's not so much about the team structure and which roles you need. It's more about the hiring process itself. Even so, I will mostly talk about data teams. Most of the principles apply to all roles, not just data.

To start with, in my opinion, hiring, especially tech hiring, is broken nowadays. I have quite some experience, both as a candidate and as a hiring manager, and I saw many things going wrong: useless and stressful interviews, lack of feedback, ghosting candidates, ghosting companies.

Sometimes you are just flooded with useless applications. For example, for a data engineering opening, I recently received a lot of CVs from junior front end developers and DevOps people, even though we explicitly stated that we are looking for somebody with at least 3 years of professional experience as a data engineer.

I know how tough the tech market is today. I totally understand desperate people who just need whatever job in tech. Nevertheless, applying for a job which obviously requires a different skill set just creates unnecessary noise.

Even though many great people are available on the market, companies keep headhunting and actively hiring people who were actually not open for jobs. So as a company, you normally have two options: either you source yourself, or maybe with your internal HR or talent acquisition team, or you go to external recruiters. And the conversion rate, interviews to offers, would probably be better. But of course, you need to pay for that service.

When I just started to hire for my teams, I was actually quite resistant to use external help. And also sometimes, even internally with HR or talent acquisition people, I was insisting on screening every single CV myself. I thought I know better whom I need.

Until I met first great recruiter who basically told me, "Ilya, just let me try. Give me a chance to provide you with five top candidates, and we take it from there." And suddenly, those were all great candidates. Even so, we didn't end up hiring some of them, but overall the amount of effort I needed to get to great interviews and great candidates in my pipeline reduced drastically.

You can go both ways. You can publish jobs on LinkedIn or other networks and go through endless candidates, or you can acquire some professional help. In both cases, it's important that you end up with decent candidates in your pipeline to have enough interviews and find the one which gets the job.

As already mentioned, market changed a lot recently. We used to headhunt candidates, we used to beg people to come to our interviews. Now it's the other way around, a lot of people on the market, but it's difficult to choose whom you talk to.

Another challenge I have: CVs are now quite generic. I suppose many people use ChatGPT or other large language models to fine-tune their CVs to match job descriptions. This leads to an interesting phenomena: companies create job descriptions using LLMs, candidates take their CVs and ask LLMs to fine-tune them and add all required keywords from the job description, and then companies need to filter through this endless LLM-generated CVs again, and mostly they use LLMs. So I asked myself, can we cut all of this and just get back to humans talking to humans?

Now let me share a few rules I created for myself after about 300 interviews I conducted as a hiring manager.

First and foremost: weak yes is a no. I think you should be afraid of false positives, bad hires. Much better to reject a great candidate compared to hiring a bad candidate. What happens if you reject somebody who would potentially be a great fit? Well, they will probably find another job, but at some point somebody will be there. On the contrary, if you hire somebody who is not a great fit for your company or your team, you risk. You risk a lot. Do yourself a favor, only hire people where you have this feeling, "yeah, this is exactly the person I was looking for."

Second rule I created for myself is: soft skills are more important than hard skills. Don't get me wrong, I'm not saying you should hire for senior roles somebody who is lacking some basic skills. Somebody who doesn't know Python for a data engineering role? Probably not. What I'm saying: yes, you have some minimal requirements on hard skills, but soft skills are so much more difficult to acquire. You can teach a person any tool or technology, but you can't teach a jerk not to be a jerk.

So again, don't risk your team, don't risk your company culture. Take a look at soft skills and attitude. And in my opinion, 80% of the success depends on those, and maybe just 20% on technologies and tools the person knows.

Talking about tools: I'm not a big fan of take-home tasks. I find it unfair and not very reliable. Unfair because the candidate spends more time and company is investing nothing. And not reliable because you basically never know who did the task. So I prefer real-time coding with a shared screen. I'm not checking algorithmic tasks or heavy, complicated things. I basically focus on communication and how people talk to me, how they approach a problem. I help them during the interview, and for me, and for my setup, it works well.

What are the differences in hiring for a data team compared to other roles? So currently, in many companies, you have cuts on budgets for data, so you have smaller teams, you have less levels of hierarchy. There is a strong need for generalists, everyone must be a bit technical and a bit business-savvy. There is no room for "it's not my job" attitude or mindset anymore.

So seek T-shaped people and not siloed specialists. The goal is to find somebody who multiplies the team value, not just adds to it.

Finalizing: hiring is still a mess, but it can be improved. Whether you're hiring or applying for a job, be yourself, be clear, be honest, intentional. Don't fake it, because it will backfire, either during interviews or at the job itself. The goal of a hiring manager should be not just to fill a role, but to find the one.

Good luck to everyone there looking for a job in data or hiring for a position in your team. I wish you good luck with that. I hope you will find your company and your candidates. Stay tuned, subscribe to my YouTube channel or LinkedIn page, and let me know in comments what you think.