RAG Q&A System

Production retrieval-augmented Q&A pipeline using LangChain LCEL, OpenAI embeddings, and Qdrant Cloud. Includes automated RAGAS evaluation, LangSmith full-chain observability, and CI/CD to AWS App Runner via GitHub Actions.

Apr 2026

AI SystemsProduction MLRAG

Overview

Production-grade RAG Q&A system with real evaluation (RAGAS faithfulness/relevancy on every query), LangSmith observability, and a CD pipeline to AWS App Runner.

Key Details

·LangChain LCEL pipeline with OpenAI text-embedding-3-small and Qdrant Cloud vector store
·0.88 average answer relevancy across test queries measured by RAGAS framework
·GitHub Actions CI/CD with Ruff + Black linting and Bandit security scanning