SELECTED WORK
A selection of research and engineering projects exploring distributed machine learning, retrieval systems, and the algorithms that make them work. Each project emphasizes its motivation, technical approach, and experimental evaluation.
PROJECT PORTFOLIO
Explore the problem, approach, and implementation behind each project.
Retrieval-Augmented Generation · Distributed Systems
A query-side embedding-drift compensation framework that uses orthogonal Procrustes alignment to recover retrieval quality across embedding-model transitions without re-embedding the entire corpus.
Federated Learning · Distributed Systems
An asynchronous federated-learning aggregation algorithm designed to maintain causal consistency independently of physical clock synchronization, using vector clocks to order updates across distributed nodes.
COLLABORATION
I'm interested in connecting with researchers and practitioners working on machine learning, distributed systems, and related areas.
Get in touch