Spec-driven development (SDD) for AI coding assistants.

Google Cloud Knowledge Catalog Tools and Samples

Developer-grade Claude Code + Codex configuration: cost-tiered subagents, workflow commands, guardrail hooks, MCP parity, and an installable plugin/marketplace.

nv-one-logger enables tracking of GPU application progress over time and can help to identify overhead from workload and cluster inefficiencies to provide efficiency metrics.

WhisperX: Automatic Speech Recognition with Word-level Timestamps (& Diarization)

Liquid Audio - Speech-to-Speech audio models by Liquid AI

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

This is a local docker container that ingrates into langchain llm just like langsmith. Provides traces for llm and LangGrah

Architecture, patterns & internals of Anthropic's AI coding agent — reverse-engineered from source maps

A Systematic Analysis and Discussion of Claude Code for Designing Today's and Future AI Agent Systems

Comprehensive reverse-engineering analysis of Claude Code's internal architecture, modules, and design patterns

A physics-grounded, cost-aware optimizer for vLLM.

kaldi-asr/kaldi is the official location of the Kaldi project.

Open Source Continuous Inference Benchmark Research Platform Kimi K2.7-Code, DeepSeekv4, GLM5 - GB200 NVL72 vs MI355X vs B200 vs GB300 NVL72 & soon™ TPUv6e/v7/Trainium2/3

very fast speech-to-text, diarization, streaming (even in CPU) with NVIDIA Parakeet in Rust

Proposals and discussions for the AI Gateway Working Group.

Skills for Real Engineers. Straight from my .claude directory.

[MLsys2026]: RAG on Everything with LEANN. Enjoy 97% storage savings while running a fast, accurate, and 100% private RAG application on your personal device.

Where the GPU for the Web work happens!

A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.

Pipecat framework based orchestrator for building real-time, voice-enabled, and multimodal conversational AI agents

TriAttention — Efficient long reasoning with trigonometric KV cache compression. Enables OpenClaw local deployment on memory-constrained GPUs.

The agent that grows with you

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