A gateway infrastructure to streamline the usage and management of various large language model (LLM) within Linkedin

Developed token and request rate limiting capability for managing access to LLMs hosted on cloud services such as Azure OpenAI Platform

An infrastructure platform that enables third party enterprise data conflation with LinkedIn

Developed an encrypted Spark-based data conflation platform, enabling third-party applications to securely combine their encrypted or hashed data with LinkedIn's dataset

Designed and developed robust APIs and mechanisms for application developers to efficiently access up-to-date data from vast datasets

Enhanced platform availability and implemented disaster recovery strategies by designing and developing failover infrastructures utilizing Azure resources

Enables offline LinkedIn feed data indexing

Collaboratively designed offline feed ingestion enhancements to increase the agility of LinkedIn's user feed infrastructure

Processed the collected metadata to generate benchmark analysis of user applications.

Processed the collected metadata to generate benchmark analysis of user applications.