Skip to main content

Productionise or Compromise!

Ai

Rewind the clock to Feb 2020 and I was returning to the Hotel on the Propel Annual Ski Trip, I’d spent the day hurtling down the mountain with little control of my speed or personal safety and was very much looking forward to 8 courses of swiss cheese and a glass of red. Sadly at dinner the conversation had taken a worrying turn – 4 British tourists had contracted coronavirus in a nearby resort and it was the first sign of things to come. 

A month later the first lockdown had started and the following weeks and months were the toughest of my career. I suddenly went from the busiest I’d ever been to having next to nothing to work on, as every client put a hiring freeze due to the uncertainty of the situation. While seeing months of work go to waste quite frankly sucked, having a clean slate allowed me the time to reflect and plan on what I wanted to do next! 

 

Something that was constantly overlapping with Python engineering and an area that a lot of my existing clients are investing in is Artificial Intelligence and Machine Learning and with that comes the requirement for talent! I have since helped a number of VC backed start ups make their first AI / ML hires in London and one issue /  question keeps popping up is the question of commercial productization and balancing that against research! 

 

On multiple occasions processes have been slowed as it hasn’t been clear in the mind of hiring managers on how much engineering experience this first data science hire has needed. On some occasions the businesses engineering teams have been able to pick up the slack and they have needed a very research focused candidate who can deep dive into the data. On other occasions they have needed very strong engineers who have some experience training models but really add value in building the infrastructure that allows these models to run in a commercial environment. And finally in some instances clients have needed someone that can do both!   

 

While finding that “Full Stack Data Scientist” will narrow the pool of candidates that will be suitable for your role it isn’t impossible to find. What can really slow down successful hiring however, is when the mix of skills you need to fit into your organization aren’t defined at the start of your search. So if at all possible before starting your hiring process make sure you have it clear in your mind whether you need this candidate to productionize or can you compromise and have your engineering team pick up the slack!

Subscribe