Reference parser and interpreter for the Dogwood policy language

From Automated Idea Factory to Realization

A browser that runs directly inside your existing terminal

Reference implementation and examples of the CuTe Layout representation and algebra.

SGLang-native serving for the Moet sign-symmetric W2 expert format with SM120 W2/W4 kernels, GLM-5.2 NVFP4 TP4 on 4x RTX PRO 6000

Stitch your rollouts together across clouds and regions

slime is an LLM post-training framework for RL Scaling.

graff — a fast agentic coding harness in Zig: multi-provider, MCP, workflows, DGM evolution loop, TS/Python SDKs

Bf-Tree is a modern read-write-optimized concurrent larger-than-memory range index in Rust from MS Research.

Phoenix is a direct hub between storage and xPU — plug in any accelerator (GPU/NPU) or AI app and stream data straight to the chip.

DeepSpec: a full-stack codebase for training and evaluating speculative decoding algorithms

Rubric compiler and judge engine for LLM evaluation

A governance harness for AI coding.

A CLI to estimate inference memory requirements for Hugging Face models, written in Python.

Programmable chat templates for LLM training and inference.

TokenSpeed is a speed-of-light LLM inference engine.

Agent operating system

Python byte transport over QUIC

A collection of DESIGN.md files analysis by popular brand design systems. Drop one into your project and let coding agents generate a matching UI.

Allow torch tensor memory to be released and resumed later

CLI for managing coding agents

Fast, accurate & comprehensive text measurement & layout

APOLLO: SGD-like Memory, AdamW-level Performance; MLSys'25 Oustanding Paper Honorable Mention

Slop Scoring to Stop Slop

PlayStation 4 emulator for Windows, Linux, macOS and FreeBSD written in C++

cuTile Rust provides a safe, tile-based kernel programming DSL for the Rust programming language. It features a safe host-side API for passing tensors to asynchronously executed kernel functions.