I lead our Core Engineering team which owns backend, cloud, data platform, devops, infra, site reliability, and security. Scaling AI-powered clinical decision support globally.

Full stack software engineering and ML engineering for AI in health care.

I worked on our ML Platform team and focused on training pipelines for large-scale recommender systems such as home timeline ranking. Elon's acquisition offer was falling apart when I left -- just in the knick of time it turned out.

I trained and deployed ML models for transaction classification for accounting and bookkeeping.

I grinded out my PhD and had a bit of fun along the way. My research focused on applied machine learning, software engineering, and human-computer interaction. Specific projects included AutoML algorithms, AutoML systems, and collaboration systems for data science projects.

I poured brews part-time for MIT students, profs, and affiliates while hacking on my research and code behind the bar. If you defended your PhD and came into the pub with your regalia, you got to sign the bucket.

I built Bayesian optimization-based hyperparameter tuning for an ML workflow orchestration product so that ML scientists could easily tune their models.

I worked on ML time series forecasting for trading volume in fixed income markets.

I worked on macroeconomic forecasting and policy analysis using the DSGE model. We were early adopters of Julia and open-sourced our entire modeling stack. I had Class I FOMC clearance and toured the gold vault during my first week. I realized I loved coding and stats but would rather read the Wall Street Journal than go to econ grad school.

Earned some extra cash during college by tutoring kids in math, science, test prep, and college applications.

I did a summer REU in the Columbia Mathematics Department working on computational problems in elliptic curve cryptography.

I worked on hard cost estimation for an affordable housing real estate developer. All the hard cost estimates in previous years had been way too low and I tried to improve the model.It later came out that the reason the hard cost estimates had been too low was because the CEOs had been fraudulently inflating the costs and splitting the money with their general contractors. A bunch of people were indicted by the feds and went to prison. This was not accounted for by my model and I'm not sure whether there are any generalizable lessons to be learned here.

I helped Professor Graciela Chichilnisky research and edit her book on climate economics focusing on carbon markets.

I was a journalist writing about companies in the Santa Barbara area and financial markets.

I assisted with the grant selection process, administrative functions, and the translation of technical summaries. I also got to go to the annual gala for free.