Mason Victors

What He Does

Recursion takes high resolution, fluorescent microscopy images of cell populations, treated with different diseases and drugs, and focuses on finding treatments for rare genetic diseases. It’s the data team’s job to extract the relevant information from these images to determine which drugs are potential treatments for which diseases. Mason coordinates projects and priorities on his data team, determining what tasks will bring them closest to achieving its goals, with respect also to the goals of the company. He also works with biologists and technicians to design the experiments to generate the most useful and valuable data. He is also responsible for improving their representation learning to extract the most useful information from the microscopy images. This involves heavy amounts of deep learning on images, while mixing in their own knowledge of biology to better inform the model representations.

Necessary Job Skills

Mason’s team is on track to generate between 10 and 20 TB of image data per week. Due to the volume, they need a way to represent each image mathematically, sometimes via deep convolutional neural networks, and to identify relationships between various drugs and diseases, which involves a whole lot of linear algebra. Add in to the mix the need for well-designed experiments (requiring fundamental statistical knowledge) and it becomes clear how critical the applied mathematics and computational skills are here at Recursion.

Pros and Cons of Her Job

Mason loves working on a problem that will change lives rather than just increasing the profits for some company. Not only is it satisfying from a societal impact perspective, it’s also incredibly technically interesting and challenging.

Work/Life Balance

As a data scientist, a lot of work can be done remotely, in addition to the collaborative atmosphere when a whole team is present. That made it easier for Mason to take paternity leave after the birth of his third daughter and to work from home two days a week when he had a two-hour commute.

Career Path

Mason bounced between physics and astronomy as a college major, but finally settled in mathematics. The only area of applied mathematics that interested him was cryptography so he planned on working for the NSA when he finished. Instead, the graduate chair at BYU convinced him to stick around and work with him on computational mathematics. During his Master’s program, Mason realized that he wasn’t really doing math, and it wasn’t really statistics or computer science either.

The term data science had only recently been coined, and it took me a while to realize that that is what I was doing. I loved it, so I embraced it completely. It was definitely accidental, but I’m grateful I stumbled into it!

After finishing his Master’s degree, Mason joined a start-up building intelligent call center software that heavily used machine learning. When it was acquired a year later, he found himself working for another data analytics company on numerous problems, ranging from natural language processing to agent-based simulation studies for workforce management. After a couple of years, he decided to get back into the start-up scene and joined Recursion Pharmaceuticals.

I think the single best career decision I made was to join a start-up right after my master’s degree. People told me “it’s risky!”, “most start-ups fail!”, “go get a Ph.D. instead and you’ll be better off in the long run!” Being in a position where you have to wear many hats at a very small start-up forces you to stretch and grow in a way that you’ll never experience in academia or at a larger company.

Career Expectations and Advice

Mason’s graduate advisor told him “The important math problems in the 20th century were largely analytical in nature. Today, they are all algorithmic” and he couldn’t agree more. While he still works through problems on a whiteboard now much like he did in college, that’s just the set-up and planning. In the end, the solutions come through computational power coupled with mathematical knowledge and modeling. Most of the time, the importance and value of answering a question is in the decision made as a result of the mathematical/computational study of data.

Sure, start-ups fail…But that doesn’t mean you fail. You learn a ton!

Salary

$100K for starting data scientists to over $250K for researchers at top institutions.