Mixture-of-experts (MoE) training megakernel for NVL72s

[ICLR'26] Stronger-MAS: A RL Framework for multi LLM agent system; [arxiv] MetaAgent-X: End-to-End Reinforcement Learning Automatic Multi-Agent Systems

PRO-V: An Efficient Program Generation Multi-Agent System for Automatic RTL Verification

MoonEP: A Perfectly Balanced Expert Parallelism Library via Dynamic Redundant Experts

AgentENV (AENV) is a distributed platform for running agent environments at scale.

An agentic system that auto-optimizes LLM workloads on AMD GPUs.

A Call of Duty-quality FPS in Three.js, built from a single prompt.

A community-maintained Python framework for creating mathematical animations.

Tenstorrent's MLIR Based Compiler. We aim to enable developers to run AI on all configurations of Tenstorrent hardware, through an open-source, general, and performant compiler.

🤘 TT-NN operator library, and TT-Metalium low level kernel programming model.

Bonsai Demo

A high-performance inference engine for AI models

Build a compiler to solve Anthropic's interview challenge.

Run GPT-5.6 Sol and Codex models inside Claude Code — private, localhost-only, one-command setup for macOS and Linux.

The fastest Mamba2 SSD kernel for TPU written in Pallas

My reasearch of losslessly compressing LLM weights.

Alex — a local LLM proxy for all your token providers, APIs and harnesses. Route Claude, ChatGPT/Codex, Gemini, Grok, Kimi & OpenRouter subscriptions into any coding tool. Your tokens, your traces,…

Use Codex from Claude Code to review code or delegate tasks.

Hyper-fast local-first LLM router: point it at your local server, burst overflow to TrustedRouter.

Anthropic's original performance take-home, now open for you to try!

🪨 why use many token when few token do trick — Claude Code skill that cuts 65% of tokens by talking like caveman

Nsight Python is a Python kernel profiling interface based on NVIDIA Nsight Tools

Machine Learning Engineering Open Book

Any model. Any hardware. Zero compromise. Built with @ziglang / @openxla / MLIR / @bazelbuild

Design visual, review-gated agent loops for Claude Code before you run them.

Skill + MCP server to turn your agent into an RLM. Load context, iterate with search/code/think tools, converge on answers.