A curated list of battle-tested tools, frameworks, and best practices for building scalable, production-grade Retrieval-Augmented Generation (RAG) systems.

GPU Memory Reservation Library

Repo for HKUST ELEC5660 Course Notes & Lab Tutorial & Project Docker

Research integrity plugin for Claude Code — paper auditing, citation verification, experiment analysis, and methodology-first skills for academic workflows.

A mapping package for Livox LiDARs

The Claude Agent Skill for Terraform and OpenTofu - testing, modules, CI/CD, and production patterns

The Python framework for state-space modeling and algorithm development

AI agents running research on single-GPU nanochat training automatically

Lightweight LiDAR-Inertial SLAM system for ROS 2. A minimal, dependency-free implementation for research and education.

A LiDAR-Inertial Odometry with Efficient Uncertainty Analysis.

Systems-level optimization of LLM inference on a single 8GB GPU — achieving 5–10× throughput improvement using static and dynamic batching without changing the model.

Official code for paper "Real-to-Sim for Highly Cluttered Environments via Physics-Consistent Inter-Object Reasoning"

RuVector is a High Performance, Real-Time, Self-Learning, Vector Graph Neural Network, and Database built in Rust.

[IROS 2025] A LiDAR-inertial odometry for dynamic environments

Lidar-IMU Localization System with Prior Map Constraint and Lio Constraint

StreamingVLM: Real-Time Understanding for Infinite Video Streams

PhysX-Anything: Simulation-Ready Physical 3D Assets from Single Image (CVPR 2026)

Learn System Design concepts and prepare for interviews using free resources.

Machine Learning Systems

This project implements a modern deep-learning-GeoAgent–based seismic signal enhancement pipeline designed for research and production workflows.

Flux 2 image generation model pure C inference

Base pretrained models and datasets in pytorch (MNIST, SVHN, CIFAR10, CIFAR100, STL10, AlexNet, VGG16, VGG19, ResNet, Inception, SqueezeNet)

Linux kernel module to gather wifi statistics from peer and non-peer STAs

mHC kernels implemented in CUDA

Published on Nature Machine Intelligence! The first real robot(quadrotor) based on differentiable physics training.