AI & ML Learning Vault
Start here: The AI ML Mind Map — three core ideas (learning is optimization, bias-variance, representation is everything), the universal ML pipeline, five problem types, and the recurring patterns that connect every technique. Read this first, return to it often.
Roadmap
Learn in order. Each section builds on the previous one.
Phase 1 — Foundations
- Math Roadmap — linear algebra, calculus, probability, statistics
- Python for ML Roadmap — numpy, pandas, matplotlib, scikit-learn basics
- Data Fundamentals Roadmap — loading, cleaning, exploring, feature engineering
Phase 2 — Core ML
- Machine Learning Roadmap — supervised, unsupervised, evaluation, tuning
- Deep Learning Roadmap — neural nets, CNNs, RNNs, transformers
Phase 3 — Domains
- NLP Roadmap — text processing, embeddings, language models
- Computer Vision Roadmap — image classification, detection, segmentation
- Reinforcement Learning Roadmap — agents, environments, policies
Phase 4 — Modern AI
- Modern AI Techniques — diffusion, RLHF, mixture of experts, agents, multimodal
- Key Papers — foundational and modern papers to study
Phase 5 — Practice
- Tutorials Roadmap — 14 hands-on tutorials, from-scratch implementations to modern AI
- MLOps Roadmap — experiment tracking, deployment, monitoring
- Projects Index — hands-on projects tying it all together
- Training Projects — train models on real datasets, from beginner to advanced
How to use this vault
- Follow the roadmap top-to-bottom
- Each note has
status:frontmatter —seed→growing→evergreen - Tag meanings:
#fundamentals#math#ml#dl#nlp#cv#rl#project [[links]]connect related concepts — use the graph view to see the big picture- Code lives in
../projects/, datasets in../datasets/