← Back to Projects Text RAG Pipeline
RAG system for document Q&A with grounded responses and clean UX.
Categories
GenAINLP
Tech Used
PythonEmbeddingsVector DBPostgreSQLPineconeRAGLLMFlaskAPIDockerGCPStreamlit
Problem
Need trustworthy answers from internal documents with reduced hallucinations.
Approach
- Designed chunking + metadata strategy
- Implemented retrieval and context constraints
- Added answer formatting and safety checks
Results
- More reliable answers
- Reusable RAG template for new corpora
Demo Videos