I have read a lot of bullshit over the years about Machine Learning Engineers or Data Scientists. They are unicorns who are perfect mathematicians, software developers, graphic design experts and strategy consultants.
I call bullshit.
So, what do we find important in finding great people to work with?
What we want to see in our ML Engineers
The most important trait of any employee. The holy grail that drives all other traits. Carol Dweck calls it Mindset, Jon Snow calls it courage, Stephen R. Covey lists it as one of the seven most crucial habits ‘sharpening the saw’.
We like employees who are interested in the world, their field, their colleagues and themselves. Employees who read articles where the only purpose is to read it for the sake of interest. Trying out new Machine Learning libraries just to see what it can do. Learning new things, just because it is fun.
Example: We have a Nest cam at the entrance of our office. One of our colleagues spent time capturing the video images, training a facial recognition algorithm and attaching a speaker. All just so you would hear “Hello Jos” the moment you first walked into the office in the morning. Useful? Not at all. Did he learn something? Definitely. Did he fail? Yes. Turns out the Nest cam API was too slow and it took about 2 seconds to get the image. By then, the employee is already 2.78 meters past the cam (assuming a leisurely walking speed of 5km/h) and doing something else.
Machine Learning is a very broad field and it is especially fast changing. There is no use in knowing things, because by the time you can claim you are a master in them, they are already old-fashioned. What is more important is being capable of quickly learning new techniques as you need them.
You are really good at building ARIMA models in R? Well, turns out Recurrent Neural Networks are better in almost every prediction task. Ow, and we train them in Python because we want to deploy them as a lambda function on AWS. We need people who can dig their way through new packages, read academic papers and implement stuff that has never been implemented before.
Example: A customer in the insurance industry asked us if we could predict the surface area of a home by using satellite photographs and frontal photographs. Whoops, even everything our computer vision expert had ever done was in classification, object tracking, regression or segmentation. She read papers on surface estimation, boundary estimation, … until she delivered a state-of-the-art model doing exactly what was asked: something innovative and new.
You created a 1000 line Jupyter Notebook solving world hunger. Great. You’re fired.
1000 line Jupyter Notebooks are useless. Code quality matters. Documentation matters. Code structure matters. Model lifecycles, and a build and release process matters. 50 years of software development best practices and methodology: they matter. Git: you matter. Not a single beginning engineer has all these skills, which is why we always pair you in an engineering squad with experienced Data Engineers or DevOps.
This is a tough one. On one side, we want you to be a code-producing, never tiring and 99% quantile productivity beast. On the other side, we want to keep our Data Scientist at our company for the next 10 years, not have them suffer from burn-outs, health issues and divorces.
At Faktion, we value your family life, your health, your happiness and your stress level. And you should value that too. We like seeing people who balance multiple aspects of their life and have a rich and fulfilling life outside of work too. Being successful at work and being successful in your family life are heavily linked and tied to each other.
Example: Both our founders are 100-hours a week maniacs. They are work machines who don’t seem to sleep at all. But at the same time, one of them is a distinguished Alpine climber, an ex-rally pilot, a family guy. When he is not working, he enjoys life to the fullest and shares that with his family, his friends and his dogs. The other one is a DJ, organizes parties and is an avid cook and general bon-vivant. He is a maniac in work, but he is also a maniac in his family life, the energy he spreads, and in maintaining the relationships with colleagues and family who he values so much.
Humans, not robots
One of the main goals of Faktion is using Artificial Intelligence and Machine Learning to automate the world and improve decision making. We create robots, thinking machines so our employees don’t have to. The human aspect of being an employee is crucial. Good data scientists can talk to other people, are not afraid to ask questions and can build relationships with their colleagues and customers.
Having great ideas and implementing awesome and innovative models is useless if you can not communicate your brilliance with the rest of the world.
Recognize yourself in this? Don’t hesitate, we need you. Fast. We are looking for great people to join the pack and run with the wolves. Check out our job openings.