Augmenting doctors with AI & Machine Learning to improve clinical outcomes and efficiency

Ran a team to rapidly expand our algorithmic surface area deploying 30+ models into new facets of the business in a 3.5 year period. This team leveraged the investments made in the data platform from my previous role

Designed the data platform & infrastructure to enable data scientists do research and deploy models to power the business. The platform at it's peak supported 150 data scientists and 80+ separate models and data products.https://algorithms-tour.stitchfix.comThis involved:- S3-based Hive data warehouse- Spark + Trino data processing- Custom ETL & Model training framework - Model Inference deployment framework- Streaming featurization framework- Experimentation platform- Observability tooling- Analytical dashboarding tools

Worked with Java, Ruby, Python, and Javascript to make Data Science using Hadoop and Apache Pig as easy as launching a Rails app on Heroku. In my time at Mortar I worked on a recommender for GitHub, added Jython UDFs to Apache Pig, contributed to Netflix’s Lipstick (the hadoop job visualizer), and most notably designed and built Watchtower.

During my two years with CUTC, I was responsible for creating and procuring all the technology for a 600 person conference running simultaneously in Vancouver and Toronto. This involved managing a team of 8 to create and design a website, registration system, and iPhone and Android companion apps.

Worked with various Microsoft technologies to a build 3D visualiza- tion of warehouse pallet loading. The software was used to help visualize a stacking algorithm for automated out-bound shipping. The project was part of a feasibility study that led to a multi-million dollar project and a patent.