Reinforcement Learning Researcher & Engineer
Haoxing Tian is a reinforcement learning researcher and engineer based in Irvine, California. He earned his Ph.D. in Electrical and Computer Engineering from Boston University, where he was co-advised by Ioannis Ch. Paschalidis and Alex Olshevsky, and received his B.Eng. in Automation from Tsinghua University.
His work bridges reinforcement learning theory and real-world systems, spanning robotic manipulation, LLM post-training, and continuing decision-making. He studies the stalibity, efficiency, and interpretability of learning systems through the lens of optimization and control theory. His research has led to first-author publications at ICLR, NeurIPS, and ICML, and he has gained industry experience at EmergenceAI and ByteDance.
I am actively seeking Research Scientist, Research Engineer, and Applied Scientist roles in reinforcement learning, robotics, and LLM systems. I am also open to postdoctoral opportunities.
Reliable and Interpretable Reinforcement Learning: Average-reward learning, finite-sample convergence, and stable policy optimization.
Robotics and Control: Continuing manipulation, efficient PPO training, and stereo perception integrated with robot control.
LLM Post-Training and Agents: Supervised fine-tuning, QLoRA, GRPO, and learning through agent memory.
I welcome conversations about research and engineering opportunities in reinforcement learning, robotics, and LLM systems.
Email: tianhx [at] bu [dot] edu
My Github, Google scholar
Resume available here (last update: 10/2026)