War Stories From The Data Science Job Search

Most people would avoid the hiring process altogether if they could. Great stories come out of job searches and I have a few to tell with lessons for capturing top talent after each one.

Vin Vashishta | Originally Published: May 20th, 2021

Book time with me for Career Coaching or sign up for my Business Strategy for Data Scientists class.

In my Data Science Recruiting Strategies class, I don’t have time to tell all the great stories of hiring process calamity I have picked up along the way. I do talk about the secondary talent market that was built so we could bypass much of the disfunction. It’s really the stories of people who tried to get hired through the traditional process that drive home why the secondary market was built and why high-end talent get their jobs that way.

 

I had one Machine Learning Engineer tell me, “How am I supposed to fit all the skills I use to architect a MLOps platform on a resume? I had a recruiter once ask me for an updated resume. I sent them a free copy of my book. What do you want me to do?”

After a certain point, people have a body of work not a resume. Most recruiters don’t expect to be talking with an author or someone with peer reviewed publications. However, in Data Science, especially with senior level talent, it is pretty common to have published.

The traditional hiring marketplace doesn’t have a way to manage the shift from analyzing a single document to an online, publicly available body of work.

 

I hired an Applied Researcher who thanked me after the phone screen for not asking any stupid questions about background type information. I said, “You’re an associate professor at {well known University}. You’ve published. What else do I need to hear?”

They told me a story about being interviewed for a Senior Research position. Part way through the interview, they realized no one on the interview team was a Researcher. No one was qualified to assess their capabilities. “Why are you going through these questions if you don’t completely understand my answers?”

“HR and the CTO told us to build a list of questions and search for answers online.”

Companies are hiring their first Researcher or Machine Learning Engineer. They often don’t have anyone who can make a qualified hiring decision. Many people avoid being the first senior level hire because they aren’t willing to navigate a hiring process that doesn’t make any sense.

 

A Senior Data Scientist told me about an interview they were in. They got asked a question about a newer model architecture. They answered. The person who asked the question disputed their answer citing the original paper about the model architecture. The Senior Data Scientist had them bring up the actual paper and it got quiet when the interviewers realized they were talking to a co-author.

I have a lot of stories like that one. Smart people who are excellent at their jobs are not always great interviewers. They’ll make up questions that sound good but don’t really assess the candidate’s ability to do the job. When that happens to senior level talent, they walk away.

In many cases, the interview process was built to assess entry level talent. Those do not translate well to high end talent and often reveal weaknesses in the team that the senior level hire is being brought in to fix.

 

One fresh college graduate told a story about an interview where they got asked a difficult question. After they finished their answer they asked, “Did I go through that completely and correctly?”

One interviewer said yes and another immediately contradicted them. They got into a heated argument over the right answer. Everyone just sat there and let it play out. What do you do to salvage an interview where people start yelling at each other?

Similar story. Multi-round interview started with the team and ended with the Director. The Director spent 20 minutes bad mouthing the team. They talked about who they were thinking of letting go. Who accepts a job offer after that?

Similar story++. One company was laying off part of the team but forcing the team to interview their replacements or they wouldn’t get their severance. Who said that out loud and still thought it was a good idea?

Companies who need someone experienced to come in and right the ship have their issues revealed during the interview process, although not normally this spectacularly. Many businesses are hiring leaders to rebuild the team around value creation and improved maturity. A traditional hiring process will drop hints about the worst of the team before the candidate can be sold on turning the team around.

 

War stories are funny after the job search is over but infuriating for candidates at the time. Senior Machine Learning talent have other options and quickly nope out of broken processes. They quickly fall into the secondary job market where new roles come to them. As soon as the community gets to see their capabilities, they are off the traditional job market for good.

With painful experiences like this one, who can blame them?