Eric Wright, Ph.D. CSEM ‘15
Microsoft Corp. Data Scientist
Former Supervising Professor: Bob Moser
What inspired you to a career in computational science?
I came to computational science via physics. I got an undergraduate physics degree, but felt like I didn't quite belong in that field. I found myself more drawn to the mathematical frameworks and computational aspects of the field than to the physical laws themselves. As I was researching graduate schools, I was thrilled to find that there was a vigorous field of study in the intersection of science, computing, and math. Computational science felt like a destiny. I was looking for it before I knew it existed.
Describe computational science to a non-expert.
In my view, science is a process for acquiring, organizing, and evaluating knowledge. Computational science extends that process to complicated scenarios beyond the reach of paper-and-pencil analysis, where millions, billions, or more, individual calculations may be required. For instance, the fundamental physical laws governing an airplane wing are (relatively) easy to write down. But how might we predict a threshold of force that damages the wing, or a find an optimal shape or material composition without actually constructing aircraft and performing numerous expensive experiments? This sort of challenge increases greatly as we move beyond human scales of space and time. How do we design molecules that have desired pharmaceutical or energy producing properties when our observational abilities are so limited at the atomic scale? The field of artificial intelligence furnishes many relevant examples here too. How can we train a machine to recognize a specific human face, or react precisely to spoken language? These are all engineering problems that have been, to some degree, solved through advances in computational science.
Describe your current job or research.
I currently work as a data scientist in the Cloud & Artificial Intelligence division of Microsoft. It's a customer-facing position on a team with fellow science, engineering, and computing Ph.D.s. In my three years there, I've worked on a variety of practical business problems like predictive machine maintenance, demand forecasting, and pricing analytics. Along the way, I've learned a great deal about the cutting edge in machine learning and cloud computing–which is like an amusement park for a computational scientist. I've also delved into topics that I never would have predicted five years ago, like econometrics and data warehousing. I really enjoy the variety of projects that I get to work on–and the opportunity to design and build products that have a wide reach in the business world.
What’s the most challenging part about your work?
There are a few challenging aspects of being a data scientist. One challenge is simply the velocity at which the field moves; keeping up-to-date with the latest machine learning/AI advances, the software tools, and the changing needs of customers is a full-time job in itself. Constantly learning is a necessity. The breadth of the data science field is a challenge as well. On a given day, I may need to be a statistician, an economist, a software engineer, or a personable consultant meeting with representatives from major corporations. Finally, there is the challenge of being a member of a team and coordinating our efforts toward a common goal over the span of months or years. I did not appreciate the difficulty or importance of this point as a student.
How does computational science “change the world?”
Behind just about all of our modern technological achievements lies some degree of computational science breakthrough. Weather and climate prediction; computer vision; genetic sequencing; signal processing; computer-aided design of vehicles and aircraft–this is just a short list of examples where computational science has made major contributions. To say that these technologies have changed the world is an understatement.
Where do you see the field in 10 years?
On a grand time-scale, I think that computational science will enable us to build accurate, predictive, mathematical models of more and more of our physical and digital universes. A vague prediction, I admit, but it's hard to see where we will end up!
In the short term, I expect that computational scientists will continue to find success in AI applications. The last decade has witnessed something of a revolution in our understanding and usage of deep neural networks in AI. Computational science will continue to feed the voracious appetites for machine learning in the AI research and business worlds. I admit this view is biased toward the tech industry, where I currently find myself. There are surely significant breakthroughs in the physical and life sciences, beyond my knowledge base, that will owe much to computational science.
What is a memorable experience at ICES?
My most memorable experience at ICES was studying for the prelim exams at the end of the first year. These memories are good and bad! But there was a great sense of camaraderie among us first-years for those tests; we really did make each other stronger mathematicians. I have fond memories of staying up late into the night working out Methods of Applied Math problems on the whiteboards in the ACES (now POB) lounge rooms.