A Full-Stack AI Bootcamp That Refuses to Skip the Boring Parts

What it does

This repository is the home base for a self-paced AI engineering bootcamp built around Jupyter notebooks and linked tutorials. It starts with Python, math, and PyTorch fundamentals, moves through MLOps topics like data validation, experiment tracking, and Dockerized cloud deployment, and then dives into modern LLM engineering: local models via Ollama, LangChain, RAG, fine-tuning, DSPy, and agent frameworks like CrewAI. The explicit goal is moving from tutorial demos to production-ready systems.

The interesting bit

Most LLM curricula treat infrastructure as an afterthought; this one puts data pipelines, DVC, MLflow, and AWS deployment on the critical path before the transformer hype begins. The syllabus is almost aggressively full-stack—linear regression and knowledge distillation share a roadmap, and CrewAI agents sit downstream of pandera validation. It is built on the premise that AI engineering is actual engineering, not just API wrangling.

Key highlights

• Full-stack scope: foundations, MLOps, LLM tooling, and context engineering in one continuous track
• Hands-on with both local and hosted models: Ollama, Llama, Gemma, Claude, Gemini, ChatGPT, Qwen, and gpt-oss
• Production tooling: FastAPI, Docker, AWS ECS, DVC, MLflow, and scikit-learn Pipelines
• Modern LLM stack: LangChain, LangGraph, DSPy, CrewAI, and Model Context Protocol (MCP)
• Each lesson pairs a written tutorial with a YouTube video and offers Colab-ready notebooks

Caveats

• The README is essentially a table of contents; all written tutorials link out to the author’s external academy site rather than rendering in-repo
• The relationship between the open repository and the linked “AI Engineering Academy” is unclear, so expect some content to live behind external gates

Verdict

A solid starting point for developers who want a structured, breadth-first ramp from traditional ML into production LLM systems. Look elsewhere if you need a deep reference on one specific framework or a ready-to-deploy project template.