Washington, District of Columbia, United States
Senior Machine Learning Engineer with experience in developing and productionizing large scale data pipelines for deep learning models. Hands on experience with both on-premise and cloud infrastructure deployments, with an emphasis on creating rigorous scalable solutions for complex cutting edge data problems.
Led work on porting on-premise based machine learning framework, data pipeline and model architecture to Google Cloud Platform. Designed a flexible data pipeline using Google BigQuery to build sequential features for model ingestion to train on 100-million member sequences. Designed and scaled a Google Cloud Dataflow pipeline using Apache Beam for customizable sequential label creation. Scaled pipeline to simultaneously label over 1000 distinct model prediction heads.
Worked with non-technical partner teams to train and productionize over 20 transformer models for predictive tasks across the organization leading to up to 60% increased value as compared to standard (non-transformer) models. Created a Spark data processing pipeline to decrease data processing of 100 million row dataset by over 300%. Designed Pytorch DataLoaders to batch load and train Causal Self Attention language model on on-premise GPU structure.
Graduate Student Instructor for upper level introductory mathematical statistics course STAT135 and introductory statistics course STAT20. Taught labs, and coding material for 70 students. Organized review sessions for over 200 students.
Alpha Epsilon Pi is the International Jewish fraternity, with more than 190 chapters and colonies across the United States, Canada, Israel, the United Kingdom, Austria and Australia.
Founded in 2014, Teramind is a leading, global provider of employee monitoring, insider threat detection, and data loss prevention solutions. Over 2000 organizations across the globe utilize Teramind’s platform to detect, record, and prevent malicious user behavior in addition to helping teams drive productivity and efficiency. The platform is available in both cloud-based and on-premise deployment options to meet any organization’s requirements.
CUJO AI is a cybersecurity company that aims to create a future where digital experiences are secure, private and personalized. Driven by proprietary AI-powered models and uniquely built on billions of real-world data points, the CUJO AI Digital Life Protection (DLP) Platform protects people and connected devices, creating a safer smart-living experience in homes, businesses and connected communities, while respecting privacy. Today, CUJO AI covers the largest inventory of devices in the world, with the broadest ability to detect, categorize and protect.