Waltham, Massachusetts

Earned a Bachelor of Science in Computer Science and Physics.

Jiangsu, China

Built Java ETL applications with JDBC + c3p0 connection pooling processing millions of records daily (−35% DB latency, +20% throughput), and a message-driven architecture with Apache Kafka, Airflow orchestration, Redis caching, and Elasticsearch full-text search.

New York, New York

Earned a Master of Science in Electrical Engineering, focused on big data and machine learning.

Toronto, Ontario

Migrated a monolith to AWS Serverless microservices in Python — AWS Textract in Lambda integrated with S3, SQS, and SNS for automated document extraction — cutting operational cost by 30% and processing wait time by 60%.

Toronto, Ontario

Built an enterprise LLM RAG system over 1M+ financial documents with a two-stage retrieval pipeline — Milvus (HNSW) + Gemini embeddings for fast candidate search, Cohere reranking for precision — raising relevance from ~60% to 95%+. Co-built horizontally-scalable Go services with Redis caching (sub-100ms P99 at 1000+ concurrent users on GCP Cloud Run), and led 2 interns rebuilding the Next.js/TypeScript frontend (~100 pages, 100+ components; 10s → <100ms loads).

Toronto, Ontario

Led the architectural redesign of a core Google ADK AI agent and built 20+ retrieval enhancements (metadata-filter search, reranking, structured-data search), lifting relevant-chunk recall from ~25% to 80%+. Optimized a Docling/Gemini/AlloyDB document pipeline (Prefect, Pub/Sub, Kubernetes) shipped through a Bazel monorepo, and ran Python data analysis that helped onboard a top-10 pharma client's 200k+ document dataset.