DeepSeek Harness: Everything is a Plugin.

Browser automation CLI for AI agents

Production-ready MoE load balancing via real-time expert replication

Litmus helps SREs and developers practice chaos engineering in a Cloud-native way. Chaos experiments are published at the ChaosHub (https://hub.litmuschaos.io). Community notes is at https://hackmd…

KernelBench: Can LLMs Write GPU Kernels? - Benchmark + Toolkit with Torch -> CUDA (+ more DSLs)

Agent-assisted and full-agent reproducibility package for MLSys 2026 FlashInfer AI Kernel Generation Contest submissions: kernels, agent workflows, skills, configs, writeup, benchmark artifacts, an…

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

Expert Parallelism Load Balancer

maximal update parametrization (µP)

[NSDI25] AutoCCL: Automated Collective Communication Tuning for Accelerating Distributed and Parallel DNN Training

Communication patterns for AI, built on top of NCCL device and host APIs

SpaceXAI's coding agent harness and TUI. Fullscreen, mouse interactive, extensible.

A CPU+GPU Profiling library that provides access to timeline traces and hardware performance counters.

A collection of tricks and tools to speed up transformer models

Run frontier MoE models on hardware you already own — pure C, zero deps, experts streamed from disk. Tiny engine, immense model. 🐦

Skill to give Claude Code (and any coding agent) the ability to generate beautiful and practical Excalidraw diagrams.

Generate beautiful dark-themed system architecture diagrams as standalone HTML/SVG files. Works as a Claude AI skill.

Venus Collective Communication Library, supported by SII and Infrawaves.

Agent skill for beautiful, verifiable architecture, workflow, sequence, data-flow, and lifecycle diagrams—self-contained HTML with motion and crisp export.

Generate draw.io diagrams from natural language — 11 presets (UML, SysML/MBSE, BPMN, network, C4…), 36 tools: codebase/CI/infra-to-diagram, image→editable diagram, mind maps, build-up animation, ex…

Open Machine Learning Compiler Framework

Agent for collecting, processing, aggregating, and writing metrics, logs, and other arbitrary data.

Kubernetes Operator for OpenTelemetry Collector

A high-performance distributed deep learning system targeting large-scale and automated distributed training.

ByteCheckpoint: An Unified Checkpointing Library for LFMs

Bernini is a unified framework for video generation and editing that combines an MLLM-based semantic planner with a DiT-based renderer.

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

Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.