Overview
Hi! I’m Alex, I eat, sleep and breathe data, and I’m here to solve your machine learning algorithm and infrastructure problems with accurate, scalable, and well-documented systems that will be effective and easy to maintain long after the project is over. I currently serve as CTO of a deep learning startup I founded to automatically find diseases in medical images, which was recently acquired by one of the largest medical imaging IT firms in Africa. Focusing on the developing world, where our products have the greatest potential to improve the most lives, we’ve developed a 14-disease chest X-ray algorithm that is now undergoing clinical trials in South Africa; it has already reportedly helped doctors there catch life-threatening diagnoses they would have otherwise missed. We’ve also achieved world-class results in CT lung nodule detection using a patent pending deep learning architecture we invented, and are carrying out an exciting pilot in CT pulmonary embolism detection with vRad, the largest radiology practice in the world. In the course of my work at Realize, I’ve built a distributed deep learning data pipeline and training infrastructure in AWS from scratch, in compliance with some of the world’s strictest privacy and security regulations, using Keras, TensorFlow, and Spark, as well as standard Python scientific computing libraries such as NumPy and Pandas. I’ve deployed our algorithms on-site and in the cloud using Docker and Kubernetes. Prior to Realize, I was a data scientist at McMaster-Carr, a multi-billion-dollar private industrial supply company, where I served as the technical lead for the company’s first deep learning project: a product search engine based on recurrent neural networks for the company’s customer-facing website. I was responsible for a number of other machine learning projects at McMaster, mostly focused on NLP, ranging from warehouse operations to optimizing website layout. I also designed and implemented a system for running and statistically analyzing large scale randomized A/B tests on the website using R. My educational background includes a master’s degree in data science from UC Berkeley, as well as a bachelor’s degree in economics and mathematical methods in the social sciences from Northwestern University.