Skip to content
View jianzhnie's full-sized avatar
🎯
Focusing
🎯
Focusing
141 followers · 228 following

Achievements

Achievement: Starstruckx3Achievement: Pull Sharkx3Achievement: QuickdrawAchievement: YOLOAchievement: Arctic Code Vault Contributor

Achievements

Achievement: Starstruckx3Achievement: Pull Sharkx3Achievement: QuickdrawAchievement: YOLOAchievement: Arctic Code Vault Contributor

Highlights

Block or report jianzhnie

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
jianzhnie/README.md
Typing SVG

📊 GitHub Stats

👨‍💻 About Me

I'm an AI engineer focused on building production-grade LLM systems and scalable reinforcement learning frameworks. I love turning cutting-edge research into clean, usable code.

  • 🔭 Currently working on distributed LLM training, model growth, and high-performance inference engines
  • 🌱 Exploring RL for reasoning, MoE scaling, and NPU/CUDA unified stacks
  • 💡 Believer in learning by building — most repos are educational but production-oriented

🚀 Featured Projects

Project Description
ScaleTorch 5D distributed training framework (DP, PP, CP, EP, TP) on PyTorch with NPU/CUDA support.
llm-grow Grow larger models from existing checkpoints: depth, width, and MoE expert expansion with function-preserving guarantees.
mini-sglang Lightweight educational implementation of SGLang (~4k lines) with PagedAttention, RadixCache, CUDA/NPU Graph, and OpenAI-compatible API.
mini-vLLM A compact implementation of vLLM, demystifying modern LLM serving systems.
LLamaTuner Easy and efficient finetuning pipelines for LLMs.
LLMEval A modular framework to evaluate LLMs across tasks and settings.

🛠️ Tech Stack

LLM Training & Serving

PyTorch vLLM SGLang DeepSpeed Megatron veRL HuggingFace FlashAttention FlashInfer Triton

Distributed & Systems

Ray CUDA Docker Kubernetes

API & Data Formats

FastAPI Safetensors OpenAI

Languages & Core

Python

📫 How to reach me

Pinned Loading

  1. LLamaTuner LLamaTuner Public

    Easy and Efficient Finetuning LLMs. (Supported LLama, LLama2, LLama3, Qwen, Baichuan, GLM , Falcon) 大模型高效量化训练+部署.

    Python 620 62

  2. Open-R1 Open-R1 Public

    The open source implementation of DeepSeek-R1. 开源复现 DeepSeek-R1

    Python 275 55

  3. deep-marl-toolkit deep-marl-toolkit Public

    MARLToolkit: The Multi-Agent Rainforcement Learning Toolkit. Include implementation of MAPPO, MADDPG, QMIX, VDN, COMA, IPPO, QTRAN, MAT...

    Python 170 21

  4. deep-rl-toolkit deep-rl-toolkit Public

    RLToolkit is a flexible and high-efficient reinforcement learning framework. Include implementation of DQN, AC,A2C, A3C, PG, DDPG, TRPO, PPO, SAC, TD3 and ....

    Python 9 2

  5. LLMToolkit LLMToolkit Public

    LLMToolkit is a toolkit for NLP(Natural Language Processing) and LLM(Large Language Models) using Pytorch.

    Python 5 2

  6. mini-vllm mini-vllm Public

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

    Python 10 2