I’m an incoming PhD student focusing on multimodal agent post-training, with the goal of enabling agents to perceive the world and make complex decisions more like humans.
My perspective is shaped by both research training and hands-on industry experience. I believe impactful research should not only push the boundaries of knowledge, but also solve real problems with scalable, simple, and robust methods.
Education
Ph.D.
University of Chinese Academy of Sciences (UCAS) · Beijing
2026 — Present
B.S. in Artificial Intelligence
Southeast University · Nanjing
2022 — 2026
Research
2025 — Present
Video Understanding & Multimodal Agent Post-Training
Video understanding, long audio–video reasoning, and multimodal agent post-training — teaching agents to hold hours of context and act on it.
2024 — 2025
Multimodal Representation Learning
Fine-grained vision–language representation and pixel-text alignment, down to any granularity.
2023 — 2024
EEG & Brain–Computer Interfaces
EEG-based BCI and mental-health modeling — reading signal directly off the brain.
Publications
2026
OmniReasoner: Thinking with Long Audio-Video via Native Tool Use
arXiv
OmniVideoBench: Towards Audio-Visual Understanding Evaluation for Omni MLLMs
ICLR
PixCLIP: Achieving Fine-grained Visual Language Understanding via Any-granularity Pixel-Text Alignment Learning
ICML
Claim-Level Rubric Rewards for Video Caption Reinforcement Learning
arXiv
2025
FlowPrune: Accelerating Attention Flow Calculation by Pruning Flow Network
NeurIPS
STGE-Former: Spatial–Temporal Graph-Enhanced Transformer for EEG-Based Major Depressive Disorder Detection
ICASSP
MLLM-CBench: A Comprehensive Benchmark for Continual Instruction Tuning of Multimodal LLMs with Chain-of-Thought Reasoning Analysis
arXiv
A Multi-scale Representation Learning Framework for Long-Term Time Series Forecasting
arXiv
* Equal contribution
Experience
Shopee · LLM Team, Pre-training
Large Language Model Algorithm Intern
Multimodal agentic post-training within the pre-training LLM team.
Sep 2025 — Apr 2026
Pinch · YC-backed AI startup
Machine Learning Engineer Intern
Voice cloning and text-to-speech for AI simultaneous interpretation — preserving speaker identity while keeping translation fast and accurate.
Jan 2025 — Jun 2025
Huawei · ICT
AI Engineer Intern
Multimodal large-model training and inference optimization — accelerating and migrating workloads from GPUs to Ascend NPUs.
Jun 2024 — Sep 2024