About Me
I began my Ph.D. at the Institute for AI Industry Research (AIR), Tsinghua University in 2022, advised by Prof. Yunxin Liu (刘云新) and working closely with Prof. Ting Cao (曹婷). Before that, I earned my B.E. in Electronic Engineering from Tsinghua University.
My doctoral research focuses on systems optimization for on-device AI inference, with the goal of running frontier models efficiently on resource-constrained devices. Representative work includes FlexNN (MobiCom 2024), Vec-LUT (MobiSys 2026, Best Paper Award Runner-Up), and OxyGen (under review).
An overview of my doctoral work: MobiSys 2026 Rising Stars Forum
I am currently working on embodied AI, exploring efficient inference and deployment of embodied foundation models (VLAs and WAMs), as well as self-evolving physical-world agents built on top of them.
Feel free to reach out: lixiangy22@mails.tsinghua.edu.cn.
News
- 2026/08 Embodied.cpp v0.1 released: a portable C++ inference runtime for embodied AI models (VLA, WAM, etc.) on heterogeneous robots. Paper Code
- 2026/07 Cosmos Lite released: deploy 16B/4B Cosmos 3 robot policies on a single 24GB RTX 4090 via quantized inference (acc. validated). Code Models
- 2026/07 DWI: Efficient On-Device LLM Decoding Through Dynamic Width Selection on Nested Models accepted to IEEE TMC. Congrats to Weining Chen. Paper
- 2026/06 Vec-LUT received Best Paper Award Runner-Up at ACM MobiSys 2026. Paper Blog Code Slides
- 2026/06 ActProbe: Action-Space Probe for Early Failure Detection of Generative Robot Policies released. Paper Code Page
- 2026/05 OxyGen updated: released ArXiv v2, and added PyTorch support (previously JAX-only) for on-board deployment (e.g., on Jetson AGX Thor). Paper Code
- 2026/05 EmbodiSkill: Skill-Aware Reflection for Self-Evolving Embodied Agents released. Paper Code
- 2026/04 Building Efficient Inference Systems for Resource-Constrained Edge AI Deployment (short paper) accepted to the Rising Stars Forum of ACM MobiSys 2026. Paper
- 2026/03 OxyGen: Unified KV Cache Management for Vision-Language-Action Models under Multi-Task Parallelism released. Paper Code
- 2026/03 Vec-LUT: Vector Table Lookup for Parallel Ultra-Low-Bit LLM Inference on Edge Devices accepted to ACM MobiSys 2026. Paper Blog Code
- 2025/08 An Empirical Study of LLM Reasoning Ability Under Strict Output Length Constraint accepted to EMNLP 2025. Congrats to Yi Sun. Paper Page
- 2025/07 Squeezer: Efficient Multi-DNN Inference for Edge Video Analytics via Cross-Model Scheduling accepted to IEEE TMC. Congrats to Xiang Wang. Paper
- 2024/01 Personal LLM Agents: Insights and Survey about the Capability, Efficiency and Security (survey & position paper) released. Paper Page 机器之心
- 2023/11 FlexNN: Efficient and Adaptive DNN Inference on Memory-Constrained Edge Devices accepted to ACM MobiCom 2024. Paper Code Slides
Awards
- ACM MobiSys 2026 Best Paper Award Runner-Up, 2026/06
- ACM MobiSys 2026 Student Travel Grant, 2026/05
- 清华之友-智能产业研究院清智奖学金, 2025/12
- 清华之友-济宁英才奖学金, 2025/11
- 清华大学社会工作优秀奖学金, 2021/12
- 第三十六届全国部分地区大学生物理竞赛非物理类A组特等奖（北京物理学会）, 2019/12
- 优秀艺术博物馆讲解志愿者（清华大学艺术博物馆、紫荆志愿者服务总队）, 2019/05
- 清华大学2018级新生二等奖学金, 2018/12
- 第三十四届全国中学生物理竞赛决赛三等奖（中国物理学会）, 2017/11
Experiences
Internship
- ByteDance, Big Data (Serverless/FaaS) Infrastructure Development, 2021/06–2021/09
Social Practice
- Student Association for Science and Technology, EE, THU (EESAST), 2019/07–2022/06
- 清华大学研究生社会实践中铁四局支队（蒙古乌兰巴托），2024/07–2024/09
- 清华大学电子工程系“E方水土·寻秘数据”支队（北京、上海），2019/08
Teaching Assistant
- Tsinghua University, Computer Program Design (undergraduate), 2025 Spring & Summer
- Tsinghua University, Basic Music Theory and Vocal Practice, 2023 Fall