
University at Buffalo
Research Assistant — AI Systems
Architected an LLM-powered research intelligence platform aggregating publications, grants, and academic activity across 80+ faculty members into automated AI-generated weekly digests.
Engineered identifier-first retrieval pipelines across 12 academic and federal data sources using ORCID, DBLP, and Semantic Scholar to improve entity resolution and retrieval consistency.
Designed a scalable PostgreSQL + Supabase knowledge architecture (12 tables, 6 views, 4 triggers) optimized for low-latency retrieval and structured LLM reporting workflows.
Built a modular FastAPI + Next.js AI platform orchestrating ingestion, retrieval augmentation, and schema-constrained LLM summarization for leadership-facing intelligence reports.



