I am currently a Data Science Intern at Binance, working on AI trading agents and trading algorithms. Previously, I interned with Tencent LightSpeed Studios' Singapore team.
My research interests center on world models and AI agents, with a focus on action-conditioned generation, policy learning, agent harnesses, and post-training for tasks such as robotic manipulation, trading algorithms, market analysis, and the replication of econometric studies. I also study the efficiency and reliability of vision-language models (VLMs).
I am seeking PhD opportunities starting in Spring or Fall 2027. Feel free to contact me at yichen013@e.ntu.edu.sg to discuss research ideas, PhD opportunities, or potential collaborations. Recommendations are also welcome. View CV (PDF)
Research Interests
World Models: Action-conditioned generation, action following, policy training and learning, and unified models, with applications in robotic manipulation.
AI Agents: Agent harnesses, agentic capabilities such as autonomous planning and tool use, and post-training through supervised fine-tuning (SFT) and reinforcement learning (RL), with applications in trading algorithms, market analysis, and the replication of econometric studies.
Vision-Language Models (VLMs): Visual token reduction, efficient inference, and hallucination mitigation to improve model efficiency and reliability.
Researching AI agents for finance and empirical research, including financial and econometric reasoning, tool use, and agent evaluation.
Contributing to InferenceNet through evaluation pipelines and reproducibility audits for agents that carry out empirical studies in economics and finance.
Working on efficient vision-language models through STAR-Pro, combining adaptive visual token selection with progressive refinement inside the language decoder.
Studying reliable multimodal reasoning and hallucination mitigation by reducing the dominance of language priors.
Exploring robotic world models, with a focus on action following, model evaluation, and policy learning.
BinanceJun 2026–Present
Data Science Intern · Remote
Supervisor: Harvin Wang
Contributing to Binance Ai Pro and multi-agent cryptocurrency trading workflows for market analysis, strategy reasoning, and trading-signal generation.
Supporting agent post-training through supervised fine-tuning and reinforcement learning, alongside model inference and deployment.
Working on agent harnesses that connect models with tools, execution workflows, and evaluation for financial tasks.
Tencent LightSpeed StudiosAug–Dec 2025
Data Scientist Intern · Singapore
Supervisor: Don Li
Conducted LLM post-training for game-specific language understanding, generation, and interactive decision-making.
Built multi-agent economic simulations with agent coordination, resource dynamics, and simulation workflows.
Harvard AI and Robotics Lab, Harvard UniversitySep–Nov 2025
Studied multi-objective reinforcement learning for financial agents and LLM-powered trading scenarios.
Explored multimodal visual-memory reasoning for medical models, including memory-enhanced post-training, multi-image inference, supervised fine-tuning, and GRPO.
Hugging Face