Curated collection of papers in machine learning systems

The repo is finally unlocked. enjoy the party! The fastest repo in history to surpass 100K stars ⭐. Join Discord: https://discord.gg/5TUQKqFWd Built in Rust using oh-my-codex.

A curated survey of database systems, design patterns, and architectural practices in modern AI systems including multi-agent frameworks, RAG pipelines, and LLM applications.

This repository contains the code for the ICLR 2026 paper “DASH: Deterministic Attention Scheduling for High-Throughput Reproducible LLM Training”, developed on top of the FlashAttention codebase.

A minimal yet professional single agent demo project that showcases the core execution pipeline and production-grade features of agents.

Accelerating MoE with IO and Tile-aware Optimizations

"Paper2Slides: From Paper to Presentation in One Click"

Miles is an enterprise-facing reinforcement learning framework for LLM and VLM post-training, forked from and co-evolving with slime.

UCCL is an efficient communication library for GPUs, covering collectives, P2P (e.g., KV cache transfer, RL weight transfer), and EP (e.g., GPU-driven)

Mirage Persistent Kernel: Compiling LLMs into a MegaKernel

Kernels, of the mega variety :)

CUDA Graph aware nvtx

Awesome-LLM-KV-Cache: A curated list of 📙Awesome LLM KV Cache Papers with Codes.

Distributed Compiler based on Triton for Parallel Systems

[ICLR 2025] DeFT: Decoding with Flash Tree-attention for Efficient Tree-structured LLM Inference

LightLLM is a Python-based LLM (Large Language Model) inference and serving framework, notable for its lightweight design, easy scalability, and high-speed performance.

DeepEP: an efficient expert-parallel communication library

FlashMLA: Efficient Multi-head Latent Attention Kernels

[NeurIPS 2025] Simple extension on vLLM to help you speed up reasoning model without training.

Puzzles for learning Triton

A unified inference and post-training framework for accelerated video generation.

Large Language Model (LLM) Systems Paper List