Sara Sadr Jahani
(sadrejahani)
AI / NLP Engineer — GenAI & RAG Systems
Data scientist turned AI engineer. I design, build and evaluate LLM systems — RAG pipelines, agents and graph-backed retrieval — and choose architectures by measured results, building on 7 years in NLP and machine learning and about 4 years of LLM applications.
What I work on
RAG & LLM Applications
End-to-end design, build and evaluation: chunking, embeddings, retrieval and grounded generation with citations, measured on golden question sets.
Agents & Retrieval Architecture
Choosing the right approach for the problem: vector vs. graph retrieval (FAISS, Neo4j), a custom agent loop vs. a framework (ReAct from scratch, LangChain, MCP).
From Data Science to GenAI
A solid NLP and ML foundation (transformers, SBERT, forecasting) applied to LLM systems, served through APIs with FastAPI and Docker.