RAG Resume API
I built a RAG API over my resume and project documents, using hybrid BM25 and local vector retrieval with cross-encoder reranking. I deployed the LlamaIndex and FastAPI service in Docker on Google Cloud Run.
Powers the live AI Q&A assistant embedded in this portfolio site
I indexed my resume and project documents so portfolio visitors can ask natural-language questions about my background, projects, and experience.
I used LlamaIndex with local ONNX BGE embeddings, hybrid BM25 and vector retrieval, and a local cross-encoder reranker. A free OpenRouter model produces context-aware answers without a paid embedding API.
I deployed the FastAPI backend in Docker on Google Cloud Run's free tier. I added per-IP rate limiting, an in-memory query cache, token streaming, and Google Sheets request logging for the public endpoint. The service powers the live Q&A assistant on this site.