I'm a physicist. I did a PhD on understanding how organic solar cells work, and I even made discoveries worthy of some coverage in mainstream media. I really enjoyed digging deep into the knowledge gaps we had about organic semiconductors. In brief:
- I was solving a relevant problem that helped people. Figuring out how electrons move in a solar cell is a very technical and abstract problem (not to mention pretty heavy on quantum mechanics and advanced maths). My motivation to do it was stemming both from my inherent curiosity and from the desire to make the world a better place with cheap, clean, renewable energy.
- I was using cutting edge technology. I flew across the ocean with our cryostat to use the facilities at RAL, where I was the first in the world to perform femtosecond stimulated Raman spectroscopy (FSRS) on solid state films of organic semiconductor heterojunctions. In our Montreal lab, we had new shiny instruments that weren't on the market yet. Prototypes don't always work, but when they do, they enable you to become a pioneer.
The transition year
I decided to leave the academic track and go into the industry. In retrospect, I made lot of logical moves which make it seem like I knew all along where I was going, but the truth is I doubted my path a lot. I wasn't sure I really wanted to leave academia or the field of physics, but I knew I wanted to try something else.
So I designed my transition year.
I gave myself one year (actually two, but it turned out to be only one after all) to test if I like working in the industry and if software development and data science is for me. One year to get the missing skills. One year to build my network. One year to find my dream job.
Here are some key ingredients that got me there:
- I connected with the community. Connecting with developers and data scientists helped a lot for informing my career moves and having exposure to interesting ideas and technologies. It broke my isolation and made me realize that many physicists now work in software startups, so it gave me confidence I could do it too. Python meetups like Montreal Python and PyLadies were my favourite. I got involved with the organization of PyLadies, which meant I was a point of contact with many companies that sponsored us and with many awesome ladies that were teaching cool stuff at the meetup.
- I ramped up in my spare time. I wanted to try out machine learning, natural language processing and big data but I had a lot to learn. So I followed as many relevant MOOCs as I could (my favourite being Andrew Ng's machine learning course), practiced with tutorials and personal projects (Kaggle was a good source of datasets to practice). And my spare time was the Christmas break between my PhD defence and my postdoc, the evenings, the weekends. I was overworking myself, but it was important to me to be knowledgeable and performant as fast as I could, so it was worth it.
- I created my job. Remember that bit about not being sure I wanted to leave academia? I decided to do a post doctoral fellowship that would be a collaboration between a university (software engineering dept.) and a company, ideally a nice startup with intellectually stimulating problems to solve. That way, I could oscillate between in and out of my comfort zone, learn what I like and what I don't. I found a great prof to be my mentor and a company, we sat together, I pitched the project (which they approved) and found some financial support from Mitacs. In that way, I created a job where I would learn and apply the new things I was interested in (machine learning, natural language processing, big data), would get to know the work environment outside of academia while still having one foot in a university to continue to do research (which I feared would be hard to do in a company, where R&D is mostly development and very little research).
During that exploration/transition year, I learned valuable things about what work I enjoy, how I enjoy to work and with whom.
- What : I can run a cluster and use it to get things done, but I really prefer to concentrate on doing the things (programming, cleaning data, analyzing data, generating insights from it) than managing a cluster or a database. (That's why I'm a data scientist/dev now, and not a data infrastructure engineer). Also machine learning is pretty fun, and you can get pretty good results with shockingly simple heuristics.
- How : I discovered pair programming and fell in love with it. Too bad I overlapped with the other dev for only a very short time, because I learned a lot that way. Version control (Git or Mercurial) were lifesavers, and so were unit tests.
- Who : I realized I had a preference for working with competent, collaborative and friendly people rather than with arrogant and/or passive-aggressive people (should be no surprise for many of you, but there's an overabundance of arrogant people in academia, so discovering this was an anomaly was nice). I also realized that for the rapid growth I was seeking, I needed to be the "dumbest person in the room", aka surround myself with brilliant people that are better than me at the skill I want to learn.
The data science years
I took those learnings at heart and found a job where I do what enjoy and am best at, surrounded by amazingly smart and friendly people. I'm a data scientist/dev at Shopify, and I make commerce better for everyone by giving data-powered advices to merchants. In brief:
- I am solving a relevant problem that helps people. Figuring out how people are shopping on the Internet is a rather technical problem (not to mention the Bayesian stats involved). My motivation to do it is stemming both from my inherent curiosity and from the desire to help merchants achieve their dreams of being their own boss and get their crafts out there in the world.
- I am using cutting edge technology. Most of the code I write runs on Spark, a parallel computing framework that is under active development, with awesome big paradigm shifts every now and then and the weirdest bugs ever. I'm proud that members of our data team at Shopify contributes to that evolving open source project and that we build our own ETL warehouse on top of it.
Hire a physicist
If you are hiring for your software company and you're reaching out only to people with a background in software engineering, I challenge you to cast a wider net and consider a scientist : physicist, chemist, biologist, neuro-scientist, you name it. These are people with an inherent scientific curiosity and strong analytical minds that would enrich your team.
And if you are looking for a job, we are hiring at Shopify. :-)