A compact implementation of SGLang, designed to demystify the complexities of modern LLM serving systems.

torchax is a PyTorch frontend for JAX. It gives JAX the ability to author JAX programs using familiar PyTorch syntax. It also provides JAX-Pytorch interoperability, meaning, one can mix JAX & Pytor…

Tokamax: A GPU and TPU kernel library.

Evaluating LLMs with Dynamic Data

Don't just regulate gradients like in Muon, regulate the weights too

Babysit your preemptible TPUs

Gemstones: A Model Suite for Multi-Faceted Scaling Laws (NeurIPS 2025)

An extremely fast Python type checker and language server, written in Rust.

Accelerate, Optimize performance with streamlined training and serving options with JAX.

High-performance, asynchronous Python HTTP client library designed for faster file transfers using concurrency, semaphores, and fault-tolerant features.

🚀 The fast, Pythonic way to build MCP servers and clients

Open-source framework for the research and development of foundation models.

Example agents I've built using the LiveKit Agents (https://github.com/livekit/agents) framework

⚡ Insanely fast AI voice assistant with <500ms response times

End-to-end realtime stack for connecting humans and AI

A Data Streaming Library for Efficient Neural Network Training

Voice activity detector (VAD) for the browser with a simple API

Official Repository for the ACL 2024 Paper: Unintended Impacts of LLM Alignment on Global Representation

The Social-IQ 2.0 Challenge Release for the Artificial Social Intelligence Workshop at ICCV '23

Python & Command-line tool to gather text and metadata on the Web: Crawling, scraping, extraction, output as CSV, JSON, HTML, MD, TXT, XML

A simple library for scaling up JAX programs

An Open Source text-to-speech system built by inverting Whisper.

Stanford NLP Python library for understanding and improving PyTorch models via interventions