Blogs & Notes | Adithya MD

Structured journeys

Series

Each series is a sequence of posts that build on one another. Start at Part 1.

Series Units 1 & 2 complete 10 classes

The Mathematics ML Runs On

The number system underneath every model, taught from the silicon up.

A from-scratch course on the math machine learning actually runs on, taught through hardware. Every class opens on a line of C that prints something impossible, then builds the math that explains it and proves it on a chip with no floating-point unit. Units 1 and 2 are complete: ten classes spanning the whole number system, then functions, sums, counting, and limits. Unit 3 is now underway on the exponential and the logarithm.

Floating Point Hardware Probability

Enter series →

Series Complete 10 parts

Before Transformers

How embeddings learn meaning, from first principles.

Ten posts that rebuild the core idea behind every modern language model: the meaning of a word can be captured by its position in a geometric space. Starts from raw text and ends with a working Skip-Gram training loop, with 28 interactive visualisations along the way.

LLMs Embeddings Geometry

Enter series →

Series Complete 11 parts

Pointers: Thinking in Memory

How embeddings learn meaning, from first principles.

Pointers are the part of C and C++ where most learners give up, most bugs live, and most seniors eventually prove their worth. This series takes them apart slowly.

Pointers Memory

Enter series →

Notes in isolation

Standalone posts

One-off writing. Essays, notes, occasional rants. Read in any order.

Nothing standalone yet. Standalone essays will live here once I write them.

Nothing matches that search. Either clear it, or I need to write more.

Blogs & long-form notes

Series

The Mathematics ML Runs On

Before Transformers

Pointers: Thinking in Memory

Standalone posts