DataTalksClub/llm-zoomcamp

LLM Zoomcamp — What This Project Is

LLM Zoomcamp is a free, open-source, 10-week online course that teaches you how to build real, production-ready applications powered by Large Language Models (LLMs) — the technology behind tools like ChatGPT. It's created by DataTalks.Club, a global community of data enthusiasts.

The course is hands-on: you learn by building, not just watching. By the end, you'll have a working codebase and the skills to build your own LLM-powered applications.

What You'll Learn

The course covers the full toolkit for building LLM apps, including:

• Retrieval-Augmented Generation (RAG) — a technique that lets an LLM answer questions using your own documents or data, rather than just its built-in knowledge
• Vector search — finding relevant information by meaning/semantics rather than exact keywords
• Embeddings — the mathematical representations that make semantic search possible
• AI agents & function calling — letting an LLM take actions and use tools
• Evaluation & monitoring — measuring how well your app performs and tracking it in production
• Hybrid search & reranking — combining keyword and semantic search for better results

Who It's For

The course is designed for people who learn by doing. It's useful for:

• Software engineers who want to add LLM features to real products
• Data engineers who want to understand how retrieval pipelines fit into production systems
• ML practitioners who want a structured way to evaluate and monitor LLM apps

Prerequisites: You should be comfortable writing Python code, using the command line, and have basic familiarity with Docker. No machine learning or LLM experience is required. No GPU needed — a standard laptop works. You'll spend about $1–5 in API credits if you run the code yourself.

The README puts it simply: "If you can write a Python function and have heard of ChatGPT, you have enough to get started."

How the Course Works

There are two ways to take it:

	Live Cohort	Self-Paced
Start	On a fixed schedule	Anytime
Lectures	Pre-recorded videos	Pre-recorded videos
Homework	Graded	Available, not scored
Leaderboard	Yes	No
Peer review	Yes	No
Certificate	Yes	No
Cost	Free	Free

Important: "Live cohort" does not mean live classes. All lectures are pre-recorded. "Live" means you work alongside others with deadlines, scored homework, peer review, and a certificate at the end.

Course Structure

The syllabus has 7 modules plus a capstone project:

1. Agentic RAG — Build a RAG pipeline with keyword search, then make it agentic with function calling
2. Vector Search — Semantic search with embeddings, using tools like minsearch, sqlitesearch, and PGVector
3. Orchestration — AI orchestration with Kestra
4. Evaluation — Measure retrieval and answer quality, both offline and online
5. Monitoring — Track user feedback and system health with live dashboards
6. Best Practices — LangChain, hybrid search, reranking
7. End-to-End Project — A complete example: a fitness assistant built with LLMs

There's also a workshop on data ingestion using dlt pipelines.

The Capstone Project

To finish the course, you build a complete RAG application of your own from scratch, including:

• A searchable knowledge base (choose a dataset, clean it, store it)
• A retrieval pipeline (the full RAG flow: retrieve context, assemble prompts, call an LLM, return answers)
• An evaluation process (measure retrieval and answer quality)
• A user-facing interface (like a Streamlit or FastAPI app)
• Monitoring and feedback loops

Past students have built fitness assistants, study companions, medical FAQ bots, codebase Q&A tools, and news summarizers.

Certificate

To earn a certificate of completion, you must:

1. Complete the final project
2. Peer-review 3 other students' projects with written feedback
3. Meet the cohort deadlines

Self-paced learners can't get a certificate but can build portfolio projects freely.

Community & Support

• Slack — The main hub for discussions, troubleshooting, and networking (join the #course-llm-zoomcamp channel)
• Telegram — For announcements
• Learning in public — The course encourages sharing your progress on LinkedIn, X, or a blog, which can earn bonus points

Optional Paid Add-On

AI Shipping Labs is an optional paid community that adds office hours, accountability sprints, personalized onboarding, and member-only workshops on top of the free course. It's a way to get extra support — the core course remains free.

Instructors & Sponsors

The course is led by Alexey Grigorev (founder of DataTalks.Club), Will Russell (Developer Relations at Kestra), and Timur Kamaliev (AI engineer). It's sponsored by Kestra (an open-source orchestration platform) and dlt Hub (an open-source data ingestion tool).

How to Contribute

If you find bugs in the materials or want to improve explanations or fix code, you can fork the repository, make your changes, and open a pull request.

Bottom Line

LLM Zoomcamp is a free, practical, community-driven course that takes you from LLM basics to building a production-ready AI assistant in 10 weeks — with the option of a live cohort experience, certificate, and a strong community to support you along the way.