My learning notes for ML SYS.

MoE training for Me and You and maybe other people

minimal DL library in C: 24 NAIVE cuda/cpu ops, autodiff engine, python API (ops bindings/layers/models), tensor abstraction, strides, complex indexing (multi-dim slices like numpy), computation-gr…

a teaching deep learning framework: the bridge from micrograd to tinygrad

High-Performance Implementation of OpenAI's TikToken.

Hallucinations (Confabulations) Document-Based Benchmark for RAG. Includes human-verified questions and answers.

s1: Simple test-time scaling

A course of learning LLM inference serving on Apple Silicon for systems engineers: build a tiny vLLM + Qwen.

The 100 line AI agent that solves GitHub issues or helps you in your command line. Radically simple, no huge configs, no giant monorepo—but scores >74% on SWE-bench verified!

This repository delivers end-to-end, code-first tutorials covering every layer of production-grade GenAI agents, guiding you from spark to scale with proven patterns and reusable blueprints for re…

Supercharge Your LLM Application Evaluations 🚀

A Book about Pythonic Application Architecture Patterns for Managing Complexity. Cosmos is the Opposite of Chaos you see. O'R. wouldn't actually let us call it "Cosmic Python" tho.

This repository contains a curated collection of 300+ case studies from over 80 companies, detailing practical applications and insights into machine learning (ML) system design. The contents are o…

Scalable and Performant Data Loading

A reading list on LLM based Synthetic Data Generation 🔥

Vocabulary Parallelism

Ongoing research training transformer models at scale

Pax is a Jax-based machine learning framework for training large scale models. Pax allows for advanced and fully configurable experimentation and parallelization, and has demonstrated industry lead…

SGLang is a fast serving framework for large language models and vision language models.

Fastest kernels written from scratch

Browser extension that simplifies the GitHub interface and adds useful features

Machine Learning Engineering Open Book

Tritonbench is a collection of PyTorch custom operators with example inputs to measure their performance.

Minimalistic 4D-parallelism distributed training framework for education purpose

This repo is meant to serve as a guide for Machine Learning/AI technical interviews.

CUDA Library Samples

Step-by-step optimization of CUDA SGEMM

Small scale distributed training of sequential deep learning models, built on Numpy and MPI.