About

About

I build things with math and computers!

I like to use the language and logic of probability to quantify uncertainty and frame problems. You could call me a statistician, or a Bayesian [statistician].

My interests (personal and professional) tend to revolve around the means and methods of mathematical modeling and their automation. I’m a big proponent of symbolic computation and its widespread use in research and industry.

Professionally, I work the whole gamut of design (mathematics, publishable papers, technical reports) and development (e.g. production level programming and numerical methods). My career has covered the areas of finance, transportation, energy, start-ups and academia, and most often in the research or planning phases where flexibility and foresight are truly needed.

Here are some situations in which I can be of service:

  • When all-purpose modeling libraries like statsmodels and scikit-learn are insufficient, or there’s no R package for the job.
  • When a situation calls for more than just the implementation, or computational betterment, of existing models.
  • When one needs to combine sophisticated models in non-trivial ways.
  • When models for high-dimensional, real-time and discrete processes are needed.
  • When a model needs automatic model selection beyond the basics (Lasso, Ridge, ElasticNet, etc.)
  • When a situation demands models that operate with [statistically] principled data preprocessing/filtering and orchestration, so that error and uncertainty can be assessed broadly.

Here’s a live copy of my CV.

Here are some things I’ve been involved in: