CV
Curriculum Vitae
Education
University of Melbourne - Master of Software Engineering · 2024.07 - present
Supervised by Dr. Ting Dang. Research in audio class-incremental learning.
Sichuan University - Bachelor of Engineering in Cyberspace Security · 2020.09 - 2024.06
Experience
Algorithm Intern, Xiong’an Institute of Artificial Intelligence · 2026.05 - 2026.08
Whole-Body WAM for Unitree G1: FastWAM model adaptation, 72-D physical actions, mixed-supervision flow matching, distributed training, multi-source data, VR teleoperation, and demonstration-data quality control.
AI Application Engineer Intern, KNQ Technology · 2025.11 - 2026.01
Built a real-time table-tennis AI commentary system spanning video, OCR, LLM, TTS, and RTMP output, plus structured match memory and pre-match narrative agents.
Selected projects
- π₀.₅ on Franka: DROID-style collection, LeRobot v2 conversion, full fine-tuning on 8×A100, WebSocket policy serving, and real-robot execution.
- Audio-L3A: 45.16% mAP on five-phase AudioSet-50, 2.65 points above LwF, at approximately one-ninth of offline training time.
- PocketLLM: 0.2B MoE decoder trained from scratch through pretraining, SFT, and DPO / PPO / GRPO alignment.
Technical focus
Robot learning and VLA · World-action models · Flow matching · JAX / PyTorch · LeRobot / DROID / LIBERO · Unitree G1 / Franka · DeepSpeed / DDP / SLURM · Audio continual learning · Python / Java / FastAPI / Spring Boot
个人简历
教育背景
墨尔本大学 - 软件工程硕士 · 2024.07 - 至今
导师:Dr. Ting Dang。研究方向:音频类增量学习。
四川大学 - 网络空间安全学士 · 2020.09 - 2024.06
工作经历
雄安人工智能研究院,算法实习生 · 2026.05 - 2026.08
参与 Unitree G1 Whole-Body WAM:FastWAM 模型扩展、72 维物理动作、混合监督 Flow Matching、分布式训练、多源数据、VR 遥操作和示范数据质检。
北京麒纪科技,AI 应用开发工程师(实习) · 2025.11 - 2026.01
构建覆盖视频、OCR、LLM、TTS 与 RTMP 输出的乒乓球赛事 AI 实时解说系统,并实现赛事结构化记忆和预赛备稿 Agent。
代表项目
- Franka 上的 π₀.₅: DROID 风格数采、LeRobot v2 转换、8×A100 全量微调、WebSocket 策略服务与真机执行。
- Audio-L3A: AudioSet-50 五阶段 mAP 45.16%,领先 LwF 2.65 个百分点,总训练时间约为离线上界的 1/9。
- PocketLLM: 从零训练 0.2B MoE Decoder,覆盖预训练、SFT 和 DPO / PPO / GRPO 对齐。
技术方向
机器人学习与 VLA · World-Action Model · Flow Matching · JAX / PyTorch · LeRobot / DROID / LIBERO · Unitree G1 / Franka · DeepSpeed / DDP / SLURM · 音频持续学习 · Python / Java / FastAPI / Spring Boot
