Continually Improving Your Hiring Process
Problem
We started with a small team of 10 people out of which one or two machine learning engineers. Part of our strategic investment included a budget for twenty more. This was a challenging prospect for many reasons. First off, we never before hired any machine learning engineers. Also, this is a very competitive job and at that time my company was not well known. We had to hire 20 people within a year and for the first six months, we hired no one. Then, from sixth to ninth month we streamlined our efforts on figuring out how to actually hire people.
Actions taken
We were not well known and have spent the first few months promoting our company. We organized a bunch of meetups, did blog posts, etc. hoping to make the New York tech community aware of us. The breakthrough occurred when we co-organized a meetup with another company that was already known for their machine learning work and when those attended stayed to learn about our company as well. From that moment on, we had people calling us and sending us resumes. The problem was how to convert all that interest into actual hires.
When people started to send us resumes we decided to record and measure everything. How many resumes we were getting, what was the conversion rate at any stage of the hiring process, etc. Also, when potential candidates rejected our offer we would ask them to share their experience. We tried to collect as much data as possible.
Every month we did a retrospective scrutinizing all the data we collected identifying weak spots and planning actions to be taken. We would solve one thing, while the new one would emerge. We realized that we were not competitive enough in terms of salaries and we adjusted that. However, even with the same salary range as our competitors, we had to offer our candidates an incentive to choose us over them. We were struggling for a while, but once we had that breakthrough event we were hiring five engineers per month. Eventually, we managed to hire 20 engineers as we were supposed to by the end of the year.
Lessons learned
- You should capture all data and do the regular evaluations trying to understand the root cause of the problem. In a nutshell, you should strive to continually improve. Many people I talked to rely on their subjective impressions to develop the ideal hiring process; instead we applied a data-based Agile to our hiring process.
- You have to be competitive. We didn’t manage to beat big brand companies but if we were only 20 percent below what they offer, money was not anymore a decision factor. This is the point when a value proposition comes to play. Machine learning engineers have at least ten offers, so it is your company’s story that should attract them. Our story was that we are a small startup that will allow them greater autonomy and a possibility to make a huge impact.
- Hiring experience matters. We measured the time it took from the initial contact to the moment when an offer was made and we managed to reduce that time from four to two or three weeks which was in huge contrast to a few months waiting period at big brand companies. We offered all candidates a red carpet treatment and small details difference providing them all together with great candidate experience.
Be notified about next articles from Laurent Paris
Connect and Learn with the Best Eng Leaders
We will send you a weekly newsletter with new mentors, circles, peer groups, content, webinars,bounties and free events.