Building a Reproducible RAG Pipeline
summary
Step‑by‑step build of a reproducible RAG pipeline that uses LangChain + Ollama + Hugging Face Embeddings, an S3-backed vector store and metadata time‑weighting to improve retrieval. Shows how a docs upload removed integrations, how git + lakeFS branching for data versioning lets you check commits, reproduce the bad answer locally, roll back or branch datasets, and iterate models and metadata until answers stabilize. Demo code, exact loaders, retriever settings and the commit that caused the bug are all shown.
AI-written summary · about
related talks
31mEasy RAG with LangChain4J and Docker
01h 12mBuild RAG using Llama 3.1
31mBuild an Agentic RAG system using Langchain, Ollama and Milvus
02h 47mFrom naive to advanced RAG: the complete guide
11mBuilding easily deployable RAG applications with Llamastack
01h 13mBuilding Production-Ready AI Systems in 90 Minutes
15mEvaluating RAG Model Performance Metrics, Bias, and Interpretability
34mImplementing an open-source RAG with OpenSearch