Mason Veilleux
Building Machine Learning and AI pipelines
I design and build machine learning and AI pipelines for optimizing decisions under uncertainty. I care most about ones that are transparent, trustworthy, and give decision-makers the confidence they need.
I am most excited about building and deploying statistical models to optimize decision-making alongside fostering a culture of trust and empowerment with data.
Selected work
Recent writing
Talks
Recent projects
Bayesian team-strength estimation for sports analytics. Recovers latent offense and defense parameters by season-week and returns posterior predictive distributions for game outcomes.
bid.optimizerCombines auction theory with Bayesian decision science so construction estimators can price bids at the profit-maximizing level in minutes rather than days.
Things that inspire me
- —Observational Price Variation in Scanner Data Does Not Reproduce Experimental Price Elasticities
A paper I usually reference when someone asks me why we need experimentation instead of using observational causal inference. I think it deeply reflects the purpose of experimentation.
- —The Soul of Erlang and Elixir
Elixir is an elegant interface for solving fault tolerance. The ability of the speaker to elucidate the primitives of the language and why it’s important is something I aspire to do. A good idea is also one that can be clearly explained.
- —You Can Just Do Things
You will usually hear me say, “you can just do things”. I believe that we are the common denominator in all things that relate to us which gives us the choice to make changes. If you have the will, you can just do things.
You can find more in my projects and occasional writings. I can be reached at masonjveilleux@gmail.com.