Leading eBay wide initiative working across multiple organizations - Enhancing developer productivity using Gen AI (Fine tuning LLMs and Retrieval Augmented Generation) - Improving reliability and performance of end to end continuous delivery pipeline - Stability of staging environment and test data.

-> Designed multiple machine learning projects using reinforcement learning, XGBoost and Logistic regression models. Scaled real time inference for 1 billion item pages traffic.-> Designed and implemented solutions for Staging data initiative across eBay, a critical company wide initiative to enhance developer productivity.

-> Designed and implemented a major distributed CDN and server side caching solution across platforms for eBay to improve site speed of item pages. https://patents.google.com/?inventor=ramesh+periyathambi-> Implemented code customization for improving neural net optimizers (Stochastic gradient descent, Adam and Adagrad) based on google tensorflow 2.0 open source code and influenced different optimizations for jacobians matrix computation using PFOR techniques by connecting with core Google Brain team.

High availability design for view item pages which will scale for all platforms.Designed New buyer platform architecture.Entire Light service design and development to load images faster (10-20 ms).

Design and development of 'seller other items' algorithm using hadoop data analysis.Whole design and core components development of 2 major domain services.

Responsible for design, development, tuning and maintenance of View Item page which is one of the critical pages for eBay. Highly involved in 2 major item page redesigns and platform changes.

Worked as senior Java/J2EE developer with various clients like Gap Inc, Kaiser Permanente IT and Screen Actors Guild