Everstore team - Meta's Binary Large Object (BLOB) storage team, responsible for serving photos, videos and files across Facebook, Instagram, WhatsApp and other products

• Architected and built DBaaS layer to provide CRUD and data versioning APIs. Achieved zero-downtime for schema transformation operations

• Developed containerized RESTful services in an event-driven and microservice based architecture

• Engineered infrastructure to store and transcode video streams, and generate real-time transcriptions

• Utilized GPU machines to accomplish 5x faster transcoding and up to 3x lower CPU usage under heavy load

• Conceptualized and released python library to improve engineering productivity of the team

• Wrote unit-tests with over 80% coverage and integrated test-suite with CI/CD pipeline

• Designed and developed language detection and geolocation identification components for Sprinklr's ML Engine which processes around 500 million documents daily

• Extensively used skills in natural language processing, machine learning, and software engineering to write performant programs

Built Smart Identity and Access Risk Analyzer. It analyzes identity, access, and activity data of employees to predict risks by using machine learning and customized algorithms.