I’ve developed skills to motivate teams to push their creative limits and give life to ideas and ways of thinking that wouldn’t exist otherwise. It’s not about having the best ideas, it’s about helping teams find them.
I believe that balance, modularization, and beauty are not only important in good technology, but are also essential in communicating new ideas that don’t yet exist to get them built.
I prefer to rapidly iterate on ideas using state of the art technologies leveraging data at scale with machine learning. However some of the best ideas are still found finding new ways to combine existing technologies in ways you didn’t expect.
Squarespace / Machine Learning Team Lead / 2012-Present
Discovering ways to leverage our extensive data at Squarespace to create a better product experience for our customers blending applied machine learning and research. We currently focus on representation learning in both the NLP and Computer Vision domains to make recommendations and predictions.
GetGlue / Sr. Software Engineer / 2011-2012
Performed a wide variety of tasks either improving existing or creating new features of the product to run on Amazon's cloud infrastructure to handle the "thundering herd" problem. Projects included topic trending detection and forecasting, who to follow recommendations, check-in recommendations, spam comment filtering, image quality processing improvements, and complex sticker unlocking dynamics.
L-1 Identity Solutions / Sr. Software Engineer / 2007-2011
Redesigned their Biometric Search Engine to a truly distributed event based asynchronous architecture that dynamically optimized complex transactions and partitioned blocking dependencies. This led to a system that automatically minimized latency while maximizing both concurrency and throughput. With the addition of adaptive scaling for different transaction types the infrastructure became part of some of the largest biometric systems in the world with efficient scaling characteristics.
Partsearch Technologies / Software Engineer / 2004-2007
Modernized external client integration from brittle batched daily ETL to an iterative real-time messaging infrastructure. This improved both data ingestion latency, and increase error resilience and data availability. Changed internal data analysis from ad-hoc queries against a denormalized database to a standardized olap data warehouse.
Keras / Tensorflow / Scikit Learn
Representation Learning / Computer Vision / NLP
Distributed Computation and Data
Spark / Hadoop / Elastic Search / Mongo / AWS
Docker / Kubernetes
REST / Swagger / Prometheus
Arduino / Raspberry Pi / Kinect