by Roy Lee

Python → AI: 1‑Year Roadmap

Goal: In 12 months, go from Python beginner → building and deploying your own ML/DL models. Study pace assumes ~6–8 hrs/week.

<aside> 💡

Study Rhythm (each week)

• Concept (60–90m): read/mini‑lecture + notes.
• Code‑along (60m): short guided notebook/script.
• Drills (60–90m): focused exercises.
• Mini‑project (90–120m): apply to a small dataset/task.
• Mastery check (20m): 6–10 Q quiz or rubric.
• Reflection (10m): What worked? What was hard? </aside>

<aside> <img src="/icons/die1_gray.svg" alt="/icons/die1_gray.svg" width="40px" />

Phase 1 — Python Foundations (Weeks 1–12)

• W1. Syntax, variables, numbers, strings, input/output, f‑strings, breakpoint().

W2. Booleans, comparisons, if/elif/else; basic arithmetic & precedence.

W3. Loops (for/while), range, enumerate, early exits.

W4. Functions: parameters, returns, scope, docstrings; simple testing with assert.

W5. Built‑ins & collections: list, tuple, dict, set; list/dict methods.

W6. Comprehensions, unpacking, *args/**kwargs, lambdas.

W7. Files & paths: read/write, CSV basics; exceptions & error messages.

W8. Modules/packages, imports, project layout, __main__ guard.

W9. Iterators/generators, yield, lazy eval; intro to time/datetime.

W10. OOP basics: classes, __init__, methods, __repr__.

W11. Testing & quality: pytest, type hints, mypy, formatting with black.

W12. Algorithms & Big‑O intuition: sorting, searching, dict/set performance.

Milestone A (end of W12): CLI mini‑app (e.g., expense tracker or quiz app) with tests.

</aside>

Phase 2 — Data & Math for ML (Weeks 13–24)

W13. NumPy: ndarrays, shapes, dtypes, broadcasting, vectorization.

W14. Pandas: Series/DataFrame, indexing, cleaning, joins, groupby.

W15. Visualization: Matplotlib basics; readable charts; saving figures.

W16. Probability & stats for ML I: distributions, sampling, confidence intuition.

W17. Probability & stats II: correlation vs causation, hypothesis testing (intuitive).

W18. Linear algebra for ML: vectors, matrices, dot product, norms; with NumPy.

W19. Calculus for ML (practical): gradients, numerical differentiation.

W20. Optimization intuition: gradient descent, learning rate, convex vs non‑convex.

W21. Data prep: feature engineering, scaling, pipelines, leakage avoidance.

W22. Model training 101 with scikit‑learn: train/test split, CV, metrics.