Xijie Huang
Research Engineer at Meta
Pittsburgh, Pennsylvania, United States
Summary
🤩
Rockstar🎓
Top SchoolXijie Huang is a research engineer focused on algorithm–hardware co-design for efficient deep learning, currently at Meta with nine years of experience bridging research and product-ready systems. He holds a PhD from HKUST and has interned at Snap and Microsoft Research, where he shipped SnapGen—an ultra-efficient T2I model that uniquely demonstrated 1024×1024 image generation on mobile in ~1–2 seconds—and developed RL-based context pruning methods for LLM few-shot learning. His background in electronic and information engineering (Shanghai Jiao Tong, top of class) informs a practical, systems-first approach to model compression and on-device inference. Xijie blends academic rigor with applied impact, repeatedly pushing model efficiency limits to enable high-resolution, low-latency AI on consumer hardware.
9 years of coding experience
Doctor of Philosophy - PhD, Computer Science Engineering (CSE), Doctor of Philosophy - PhD, Computer Science Engineering (CSE) at 香港科技大学工学学士学位, 电子信息与电气工程, 89.4/100 (91.3/100 for junior year) Ranking: 2/55, 工学学士学位, 电子信息与电气工程, 89.4/100 (91.3/100 for junior year) Ranking: 2/55 at 上海交通大学
Chinese, English
Github Skills (33)
agent10
action-recognition10
lora10
quantization10
peft9
transformers9
llm9
large-language-models9
llama9
fine-tuning9
mistral9
rlhf9
ai9
mask9
gpt9
Programming languages (3)
Github contributions (5)
HuangOwen/Awesome-LLM-Compression
Jun 2023 - Apr 2025
Awesome LLM compression research papers and tools.
Contributions:4 reviews, 34 PRs, 455 pushes in 1 year 10 months
HuangOwen/HuangOwen.github.io
Nov 2018 - Feb 2025
Contributions:109 pushes, 1 branch in 6 years 4 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.