jamwithai/production-agentic-rag-course is a learner-focused project for building a complete, production-grade AI research assistant step by step. It emphasizes search fundamentals before adding vector-based hybrid retrieval.
Project overview
It presents a professional learning path from keyword search foundations to hybrid retrieval, then extends the assistant with local-LLM RAG, monitoring, and agentic workflows.
Project type
AI Agent · RAG · Infrastructure
Use cases
Knowledge Q&A · Learning & Education
Deployment
Refer to project documentation
License
MIT
Best for
Learners who want step-by-step guidance for building a production-grade AI research assistant grounded in search fundamentals.
Key capabilities
Sets up infrastructure with Docker Compose, FastAPI, PostgreSQL, OpenSearch, Airflow, and a local Ollama LLM.
Provides an automated pipeline that fetches and parses academic papers from arXiv using Docling.
Provides production BM25 keyword search in OpenSearch with filtering and relevance scoring.
Combines intelligent chunking with hybrid retrieval that merges keyword and semantic search using Jina AI embeddings.
Provides a RAG pipeline through a local Ollama LLM, with streaming responses and an interactive Gradio web interface.
Adds RAG monitoring and tracing with Langfuse, plus Redis caching for optimized performance.
Implements a LangGraph agentic workflow with decision-making, document grading, query rewriting, and guardrails.
Provides mobile access and conversational AI for the research assistant through a Telegram bot.
Limitations and risks
Requires at least 8GB RAM and 20GB of free disk space.
Requires Docker Desktop, Python 3.12+, and UV Package Manager to be installed.
Getting started
Install Docker Desktop, Python 3.12+, and UV Package Manager; clone the repository, change into it, copy `.env.example` to `.env`, run `uv sync`, then run `docker compose up --build -d`.
Evidence and sources
README: Master the most in-demand AI engineering skills: <strong>RAG (Retrieval-Augmented Generation)</strong>
README: This is a learner-focused project where you'll build a complete research assistant system
README: The arXiv Paper Curator will teach you to build a production-grade RAG system using industry best practices. Unlike tutorials that jump straight to vector search, we follow the pr…