#3131st of 118 in Learning and cookbooks

artificial-intelligence-deep-learning-machine-learning-tutorials

Self-checking Jupyter lessons that build AI components from scratch

Stars
4k
License
custom license
Last commit
Nov 2026
Last release
Nov 2026
Language
Python
Stars, last 29 weeks+74
25 Apr 2026: 3,953 stars2 May 2026: 3,953 stars9 May 2026: 3,955 stars16 May 2026: 3,958 stars23 May 2026: 3,960 stars30 May 2026: 3,963 stars6 Jun 2026: 3,966 stars13 Jun 2026: 3,968 stars20 Jun 2026: 3,970 stars27 Jun 2026: 3,971 stars4 Jul 2026: 3,972 stars11 Jul 2026: 3,974 stars18 Jul 2026: 3,975 stars25 Jul 2026: 3,977 stars1 Aug 2026: 3,980 stars8 Aug 2026: 3,983 stars15 Aug 2026: 3,987 stars22 Aug 2026: 3,990 stars29 Aug 2026: 3,991 stars5 Sept 2026: 3,991 stars12 Sept 2026: 3,993 stars19 Sept 2026: 3,997 stars26 Sept 2026: 3,999 stars3 Oct 2026: 4,006 stars10 Oct 2026: 4,006 stars17 Oct 2026: 4,009 stars24 Oct 2026: 4,016 stars31 Oct 2026: 4,018 stars7 Nov 2026: 4,026 stars14 Nov 2026: 4,027 stars

3,953 on 25 Apr 20264,027 on 14 Nov 2026

From GitHub's daily star counts.

Checked on GitHub.

Overview

Synapsa Commons is a set of 103 Jupyter notebook lessons that build AI components from scratch, such as autodiff, BPE tokenizers, a CPU-trained language model, GRPO, retrieval and an agent loop. Each exercise checks itself when run. Lessons are grouped into tracks and domain programmes (EU AI Act conformity, document intelligence, model risk) and open in Colab, Kaggle, Binder, Codespaces or local Jupyter.

Who it is for: Engineers and learners who want runnable, hands-on AI lessons

Strengths

  • Exercises self-check when run, with hints on likely mistakes
  • Starter lesson runs on a laptop CPU with 8 GiB RAM, no GPU
  • Opens in Colab, Kaggle, Binder, Codespaces or local Jupyter
  • Covers agents, MCP, RAG, post-training and inference, plus 7 C/C++ exercises

Weaknesses

  • Two programmes are partly built: credit underwriting 4 of 8, proteomics 5 of 7
  • The 2017-2025 tutorials were retired and are no longer maintained here
  • License is detected as NOASSERTION, so reuse terms are unclear
  • Lessons are teaching-scale from-scratch builds, not production libraries

What it needs

  • RAM ≥ 8 GB
  • no GPU
  • Needs Python, Jupyter, numpy

What changed

From the release notes. No flag means the note does not say so, not that the upgrade is safe.

Also in Learning and cookbooks

See all 118
Also in Learning and cookbooks
RankProjectScore
1ai-engineering-from-scratchFree curriculum of 542 coding lessons on building AI systems from scratch79.4k starsMIT91 out of 100
2generative-ai-for-beginners21-lesson course on building generative AI apps in Python and TypeScript123.8k starsMIT83 out of 100
3rag_techniquesCollection of 42 runnable notebooks implementing RAG techniques30.4k starscustom license83 out of 100
#4cs249r_bookOpen textbook series on ML systems, with TinyTorch, labs and hardware kits30.2k starscustom license83 out of 100
#5agents-towards-productionJupyter tutorials for taking LLM agents from prototype to production21.9k starscustom license82 out of 100