Free, open-source Beamer slide decks and code for Machine Learning, Deep Learning, NLP, Generative AI, Maths for ML, and Python, from 1-hour seminars to full courses.
Everything here, including slides, code, and notes, has been built by learning from, and citing, the best public material available, and is given back under an open license so anyone can teach, learn, or build on it.
A few sample slides, actual output from the .tex sources in this repo, no editing:
 |
 |
 |
Content comes in three tiers, so you can pick exactly the depth you need:
- Courses (full curriculum, ~1-2 weeks / ~40 hrs): Machine Learning, Deep Learning, Generative AI, Maths for ML, Python
- Workshops: one topic in depth, ~1-2 days (8-16 hrs)
- Seminars: a focused session, ~1-2 hrs
Workshops and seminars span far more ground than the courses above: classical ML, linear algebra, LLMs, RAG, LangGraph, GNNs, career advice, and more. There's also a growing set of runnable code projects in Code/: LangChain, LangGraph, RAG pipelines, fine-tuning, classical ML, NLP, graph neural networks, and more.
Every seminar/workshop compiles to both Beamer slides and two-column printable "cheat sheet" notes from the same source .tex files.
Five full courses (Machine Learning, Maths for ML, Python, Deep Learning, Generative AI), each assembled from standalone workshops and seminars you can also take independently, plus dozens more standalone workshops and seminars beyond them (NLP, LLMs & GenAI, Graph ML, Reinforcement Learning, Software Engineering, Data Analytics, career prep, and more). See COURSES.md for the full catalog with links to every driver.
Code/ has runnable Python/notebook projects, organized by area:
| Category | Examples |
|---|---|
| GenAI / Agents | langchain/, langgraph/, llamaindex/, crewai/, agents/, agno/, google-adk/ |
| RAG Applications | chatbot-faqs/, chatbot-multimodal/, omni-rag/, parsing/, graphrag/ |
| LLM Fine-tuning | fine-tuning/ |
| Document Parsing | docling/, opendataloader/ |
| Deep Learning | pytorch/ |
| Classical ML | ml/, math/, python/ |
| NLP | nlp/, dnlp/, spacy/ |
| GNN | gnn/ |
| Indic Language | mahamarathi/, sarvam/, orgpedia/ |
Each project has its own environment.yml (conda) and, for most, a test_*.py suite runnable with pytest.
- Mission: Spread knowledge of Data Science to a wider audience.
- Vision: Let many participate in the industry moving "From Automation to Autonomy."
- Values: Give back, pay it forward.
This is a LaTeX repository. Slides compile to PDF; they aren't checked in pre-built. To try one:
cd LaTeX
texify -cp Main_Seminar_ML_Intro_Presentation.tex
Requirements: a LaTeX distribution (MikTeX or TeX Live); install packages as prompted. Every seminar and workshop has both a _Presentation.tex (slides) and a _CheatSheet.tex (two-column notes) driver. Compile either.
LaTeX/:.texsources and images. Naming:subject_maintopic_subtopic.tex(e.g.maths_linearalgebra_matrices.tex). Driver files:Main_[Seminar|Workshop|Course]_<Subject>_[Presentation|CheatSheet].tex. Content hierarchy: Course (40hr) → Workshop (4-16hr) → Seminar (1hr) → raw topic files.Code/: runnable Python/notebook files with necessary images and data, named to match their corresponding.texfile where applicable.- A
References/directory (papers, code, slides used as source material) exists locally but isn't uploaded, since most of it belongs to others' repos rather than original work.
See CONTRIBUTING.md: corrections, new topics, code examples, and citation fixes are all welcome.
Notes have been built from lots of publicly available material; care has been taken to cite sources, but some may be missing. Point them out and they'll be fixed. Don't depend on this fully; there's always more to improve.
Connect: LinkedIn
Copyright (C) 2019 Yogesh H Kulkarni · MIT License