Bobby Peng

Bobby(Zhuoran) Peng

I was a CS & Math undergrad & now master student at New York University's CILVR Lab, where I am fortunately advised by Prof Lerrel Pinto.

My interest lies on Robot learning, particularly lifelong learning, multitask learning, and World Models.

In my spare time I cook, eat, play tennis, and watch anime.

Research

	PROTIP - Points and Reasoning Orchestration: Tools for Integrated Policies Zhuoran Peng, Raunaq Bhirangi, Lerrel Pinto, Siddhant Haldar ICRA, 2026, under review project page / arXiv A multi-agent orchestration system for high-level planning with vision models and low-level point-based control.
	BAKU: An Efficient Transformer for Multi-Task Policy Learning Siddhant Haldar, Zhuoran Peng, Lerrel Pinto NeurIPS, 2024 poster project page / arXiv A simple transformer architecture that enables efficient learning of multi-task robot policies.
	OPEN TEACH: A Versatile Teleoperation System for Robotic Manipulation Aadhithya Iyer, Zhuoran Peng, Yinlong Dai, Irmark Guzey, Siddhant Haldar, Soumith Chintala, Lerrel Pinto CoRL, 2024 poster project page / arXiv Open-sourced vision-based robot teleoperation system for multi-fingered dexterity and bimanual manipulation, currently supporting Franka, xArm, Jaco, Stretch, and Allegro platforms
	Hyperspectral Image Recovery Constrained by Multi‑Granularity Non‑Local Self‑Similarity Priors Zhuoran Peng, Yiqing Shen Under Review, 2025 project page / arXiv Use the Multi‑Granularity Non‑Local Self‑Similarity (MG‑NSS) prior for hyperspectral image (HSI) recovery, unifying coarse‑grained Tucker and fine‑grained FCTN decompositions to jointly capture global, local and non‑local cues