# AI Bootcamp by curiousily: A Free Self-Paced Generative AI Engineering Course

> AI Bootcamp is a MIT-licensed collection of Jupyter notebooks maintained by Venelin Valkov at mlexpert.io, covering the full AI engineering stack from Python and PyTorch fundamentals through Ollama, RAG, LangGraph, fine-tuning, and multi-agent systems. It is a hands-on reference, not a structured course with grading; value comes from running the notebooks, not from reading them.

**curiousily/AI-Bootcamp** — Self-paced bootcamp on Generative AI. Tutorials on ML fundamentals, Ollama, LLMs, RAGs, LangChain, LangGraph, Fine-tuning, DSPy & AI Agents (CrewAI), (Using ChatGPT, gpt-oss, Claude, Qwen, Gemma, Llama, Gemini)

- Repository: https://github.com/curiousily/AI-Bootcamp
- Website: https://mlexpert.io
- Stars: 946 · Forks: 293
- Language: Jupyter Notebook
- License: MIT
- Published: 2026-09-10 · Updated: 2026-09-10 · Language: en
- Canonical page: https://hysenlabs.com/projects/curiousily-ai-bootcamp

## What AI Bootcamp Is and Who It Targets

AI Bootcamp is a GitHub repository of Jupyter notebooks described by its README as 'Get Shit Done with AI,' focused on real-world applications for AI engineering. It is maintained by Venelin Valkov, who also runs the AI Engineering Academy at mlexpert.io. The repository covers the full engineering stack: Python essentials, machine learning fundamentals, MLOps, and a large section on generative AI systems including Ollama, LLMs, RAG, LangChain, LangGraph, fine-tuning, DSPy, and AI agents.

The target learner is a working developer or data scientist who wants practical exposure to specific generative AI techniques. The README does not position this as an introductory course for people new to programming; the AI/ML Foundations track assumes you can follow Python code and understand basic ML concepts. More advanced tracks in AI Systems Engineering assume comfort with LLMs, APIs, and containerized deployments.

All content in the repository is free. The project is MIT-licensed. The mlexpert.io AI Engineering Academy is mentioned as a companion platform for learners who want guided instruction alongside the notebooks, but the notebooks themselves require no subscription or login.

## Repository Structure: Numbered Notebooks and Model-Specific Files

The repository organizes content into two groups. The first is a series of numbered notebooks starting with 01.python-essentials-for-ai.ipynb and progressing through 30.multimodal-embeddings-rag.ipynb. These follow a deliberate sequence: foundations come first, then MLOps topics, then generative AI systems. Running them in order gives a coherent progression from basics to multi-agent systems.

The second group is model-specific notebooks named after the models they cover: llama-3.ipynb, llama-3-3.ipynb, gemma-4.ipynb, gpt-4o.ipynb, deepseek-r1.ipynb, deepseek-v3.ipynb, gemini-3-flash.ipynb, and others. These are not part of the numbered sequence. They serve as standalone references for working with a specific model and reflect the project's approach of adding new notebooks as notable models appear.

Other notebooks cover specialized topics outside the numbered sequence: FLUX-1-Kontext-Dev.ipynb (image generation), MinerU2.5-2509-1.2B.ipynb (document extraction), ModernBERT.ipynb, and GLM-OCR.ipynb. The top-level structure is flat: all notebooks and model-specific files sit in the root directory alongside the README and pyproject.toml.

## Getting the Notebooks Running: Colab or Local Python 3.13

The README includes a Colab badge linking to https://colab.research.google.com/github/curiousily/AI-Bootcamp/, which opens any notebook in Google Colab directly from the repository. This requires a Google account but no local setup. Colab provides GPU access for free at the cost of session limits; longer fine-tuning notebooks may hit those limits.

For local use, the pyproject.toml defines the project as mlexpert_bootcamp and sets the minimum Python version at 3.13.9. The dependencies include jupyterlab (4.6.3 or later), openai (2.43.0 or later), pandas (3.0.6 or later), matplotlib (3.11.2 or later), and tools for PDF processing and other specialized notebook requirements.

The project uses a .pre-commit-config.yaml and ruff for linting, suggesting the maintainer cares about code consistency across notebooks. The .python-version file pins the Python version used in the development environment, which aligns with the pyproject.toml constraint.

## Three Learning Tracks: Foundations, MLOps, and AI Systems Engineering

The README describes three distinct tracks.

The AI/ML Foundations track covers Python essentials (data structures, NumPy, Pandas), mathematics for AI (linear algebra, calculus, probability), linear models (regression and classification baselines), PyTorch fundamentals (tensors, autograd, training loops), and practical data exploration using the Bank Marketing dataset with pandas and Seaborn.

The MLOps and Production Systems track moves into production concerns: data validation with pandera, scikit-learn Pipelines, DVC for versioning, experiment tracking with MLflow and LightGBM, building FastAPI REST APIs from trained models, Dockerizing those APIs, and deploying to AWS using ECS and EC2.

The AI Systems Engineering track covers the generative AI stack that most engineers need today: Ollama for running models locally (notebook 28), basic LLM usage (notebook 5), building RAG pipelines (notebook 6), advanced RAG with LLaMA 3 in LangChain (notebook 7), fine-tuning (notebook 8), deploying LLMs (notebook 9), agents with LangGraph (notebooks 18, 19, 21, 29), SQL agents with CrewAI (notebook 14), PydanticAI agents (notebook 24), and LiteLLM for multi-provider LLM access (notebook 26).

This breadth is a strength for exploration and a potential distraction for learners who need depth in one area. The notebooks provide working examples, not extended explanations of why each design decision was made.

## Model-Specific Notebooks: Current Models Without Version Pinning

A notable pattern in this repository is the model-specific notebook series. As new LLM releases appear, new notebooks follow: llama-3.ipynb, llama-3-2.ipynb, and llama-3-3.ipynb each cover a successive LLaMA release. Gemma gets gemma-3.ipynb, gemma-3n.ipynb, gemma4.ipynb, and gemma4-12B.ipynb. DeepSeek appears as deepseek-r1.ipynb and deepseek-v3.ipynb. The Gemini series includes gemini-2-flash-thinking.ipynb and gemini-3-flash.ipynb.

This pattern means the repository stays current with notable model releases, which makes it useful as a quick reference when a new model ships. The downside is that older notebooks reference models that may have been superseded or deprecated; a notebook for llama-3.ipynb does not automatically update when Llama 3.3 changes the recommended usage pattern.

The gpt-oss-20b.ipynb notebook, for instance, covers a specific model that may not be the current recommendation for any given use case. Learners should treat model-specific notebooks as starting points to adapt, not as perpetually current references.

## Comparing AI Bootcamp to fast.ai and What It Leaves Out

fast.ai provides a well-known free deep learning course taught through Jupyter notebooks with a top-down pedagogical approach, emphasizing intuition before theory. AI Bootcamp covers a wider surface area (from Python basics through MLOps and generative AI agents) but with less depth at any individual topic. The fast.ai course is designed to be taken in a fixed order with a specific teaching philosophy; AI Bootcamp is more of a reference library organized sequentially than a course with a pedagogical framework.

AI Bootcamp does not include exercises, quizzes, or progress tracking. There is no certificate of completion. The Discord community linked in the README and the mlexpert.io Academy provide the social and guided-learning layer, but those are external to the repository itself.

The pyproject.toml constraint of Python 3.13.9 is a real limitation for teams on earlier Python versions. Some notebooks have dependencies that are not globally installed at the project level, which means individual notebooks may fail without specific pip installations inside the notebook itself. The last push to the repository was on 2026-09-19. The project is MIT-licensed.

## Conclusion

AI Bootcamp is well suited for engineers who want self-paced, hands-on exposure to the full generative AI engineering stack, from PyTorch fundamentals to LangGraph agents, using up-to-date model-specific notebooks. It is not a structured course with quizzes, certificates, or a syllabus: learners who need accountability or formal credentials will find it thin on those fronts. Before starting locally, verify that your environment runs Python 3.13.9 or later; the pyproject.toml sets this as the minimum version, and notebooks will fail on older Python versions.

## FAQ

### What Python version does AI Bootcamp require?

The pyproject.toml sets requires-python to 3.13.9 or later. Running notebooks on earlier Python versions is not supported by the project's configuration. Using Google Colab bypasses local version constraints, though Colab's Python version may differ from the pinned version.

### Does AI Bootcamp require API keys for the LLM notebooks?

The pyproject.toml includes openai as a dependency, and the notebook topics cover OpenAI, Anthropic (Claude), Gemini, and open-weight models via Ollama. Notebooks that call cloud LLM APIs require the corresponding API keys. Notebooks that use Ollama can run without external API access by running models locally.

### Is AI Bootcamp related to the mlexpert.io AI Engineering Academy?

Yes. The README links to the AI Engineering Academy at mlexpert.io as a companion platform offering guided instruction. Each lesson row in the README table links to a tutorial on mlexpert.io. The GitHub repository is the open, free layer; the Academy is a separate subscription-based offering.

## Sources

- [curiousily/AI-Bootcamp on GitHub](https://github.com/curiousily/AI-Bootcamp)
- [Issues](https://github.com/curiousily/AI-Bootcamp/issues)
- [License: MIT](https://github.com/curiousily/AI-Bootcamp/blob/master/LICENSE)
- [Project website](https://mlexpert.io)
- [README](https://github.com/curiousily/AI-Bootcamp/blob/master/README.md)

---

Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/curiousily-ai-bootcamp