Workspace Learning Review

Audit of the repo as a learning system. Main risks at this scale: navigation drift, stale guidance, uneven scaffolding.

Key Findings

  1. Several phase READMEs described wrong file layouts or old phase numbering.

  2. Root-level entry point pushed learners toward a weak or nonexistent orientation flow.

  3. 23-glossary/ and 28-practical-data-science/ were missing usable landing pages.

  4. 28-practical-data-science/00_START_HERE.ipynb was empty.

  5. Broken markdown links in top-level and phase-level guides.

Folder Priority

Priority

Folders

High

00-course-setup, 01-python, 05-embeddings, 06-neural-networks, 11-prompt-engineering, 12-llm-finetuning, 14-local-llms, 23-glossary, 28-practical-data-science

Medium

02-data-science, 03-maths, 04-token, 07-vector-databases, 08-rag, 09-mlops, 10-specializations, 20-real-time-streaming, 21-quizzes, 24-advanced-deep-learning, 25-reinforcement-learning, 26-time-series, 27-causal-inference

Low

15-ai-agents, 16-model-evaluation, 17-debugging, 18-low-code, 19-ai-safety, 22-references, scripts

Fixes Applied

  • Corrected repo entry-point guidance in main README and setup docs

  • Fixed stale links in 01-python/README.md

  • Rewrote phase READMEs for embeddings, neural networks, prompt engineering, fine-tuning, and local LLMs to match actual notebooks

  • Added landing pages for 23-glossary/ and 28-practical-data-science/

  • Added real start notebook for 28-practical-data-science/