End-to-end RAG pipeline: chunking + embedding ingestion, vector index with metadata filtering, hybrid retrieval, and streamed LLM responses with citation rendering. Tuned chunk strategy and re-ranking for measurable improvements in answer accuracy and reduced hallucinations on a domain-specific corpus.
Back to selected work
② Case study/projects/rag-app
RAG-Based AI Application.
Retrieval-Augmented Generation system built with embeddings and vector search. Implements semantic search, streaming AI responses, and a retrieval pipeline tuned for answer accuracy and relevance.
StackSvelteKitOpenAIVector DBTypeScript
01 / 01
↳ Keep browsing
See more recent runs.