Research Engineering: distributed compute framework that empowers high velocity research

Platform Engineering: tooling and abstractions over modern day technology to enable research and development

Stack: python, kubernetes, docker, grpc, gcp

Research & Trade Analytics Platform

Options Market Making Technology

• Implemented centralized Redis feature store which reduced memory usage on 52 EC2 instances by 75%

• Reduced initial request latency by 90% by caching model weights and multithreading feature computation

• Added new features to the product ranking model to increase GMV (gross merchandise value) by >1%

• Spearheaded the development of the Asset Management Service for the launch of Watson Studio Desktop

• Architected file system utility APIs that made up 60% of the backend for the critical service

• Developed a custom document indexer in order to perform complex search queries using Apache Lucene

• Wrote and maintained backup scripts that ensured the resiliency of user data in Watson Knowledge Catalog