plan.md

March

I'm going to be splitting my time with 6 hours a week on iOS dev, 4 hours on Rust and 2 on ML.

Weekly tasks for March

iOS development

Continue to progress through 100 days of Hacking with Swift.

Week 1 ✅

Rust - Ownership & Moves ✅

Implement a CLI tool:

• Reads file
• Counts word frequency
• Outputs sorted results

Constraints:

• No cloning unless necessary
• Explain every move

Questions:

• When does a move occur?
• Why does String move but i32 copy?

ML - Linear Algebra ✅

Study:

• Vectors
• Dot product
• Matrix multiplication

Implement in Python:

• Basic vector ops from scratch (don't use numpy aha)

Week 2

Rust - Borrowing & References

Refactor week 1 tool to:

• Avoid unnecessary moves.
• Use references.
• Implement functions that borrow instead of take ownership.
• Intentionally create borrow checker errors, then fix them.

Document:

• Why the error occurs.
• What rule is being enforced.

ML - Cost Function

Derive: $$ J(0) = \frac{1}{2m}\sum (h_{\theta}(x) - y)^2 $$ Explain:

• Why square the error?
• Why divide by m?
• What shape are these objects?

Week 3

Rust - Lifetimes

Create:

• A small struct holding references.
• Functions returning borrowed data.
• Add explicit lifetime annotations.

Explain:

• What lifetime annotations actually mean.
• Why Rust can't infer some cases.

ML - Gradient Descent

Derive update rule: $$ \theta \colonequals \theta - \alpha \nabla J(\theta) $$ Implement:

• Gradient descent loop manually.

Week 4

Rust - Intro to Concurrency

Implement:

• Multi-threaded word counter using Arc + Mutex Questions:
• Why is shared mutable state dangerous?
• What problem does Mutex solve?
• What is a data race formally?

ML - Full implementation

Combine:

• Synthetic dataset
• Manual regression
• Plot loss decreasing

Rules:

• No ML frameworks
• Optional: NumbPy for convenience only after manual implementation