Juan Carlos Niebles

Research Director, Salesforce AI Research
Co-Director, Stanford Vision and Learning Lab
Adjunct Professor, CS Dept., Stanford University

picture.jpg

bio

Juan Carlos Niebles is a Research Director at Salesforce and an Adjunct Professor of Computer Science at Stanford University, where he serves as co-Director of the Stanford Vision and Learning Lab. His research focuses on the intersection of computer vision, machine learning, multimodal AI, and autonomous agents.

With over 100 published articles in top-tier venues, Juan Carlos is a recognized leader in the global AI community. His career spans significant industry-academic leadership, including roles as Associate Director of Research for the Stanford-Toyota Center for AI Research, and Senior Research Scientist at the Stanford AI Lab. He also held a long-term professorship at Universidad del Norte in Colombia.

Beyond research, he helps shape the AI ecosystem as a member of the AI Index Steering Committee, Curriculum Director for Stanford-AI4ALL, and as an Area Chair for CVPR, ICCV, and ECCV. He also served as an Associate Editor for IEEE TPAMI. His contributions have been recognized with the Microsoft Research Faculty Fellowship, several Google Research Awards, and a Fulbright Fellowship. In 2025, he was named one of the Top 100 most prominent leaders in AI in Colombia and was previously a Forty Under Forty recipient.

He holds a Ph.D. degree in Electrical Engineering from Princeton University, an M.Sc. from the University of Illinois at Urbana-Champaign, and an Electronics Engineering degree from Universidad del Norte.

research

My research is centered on Computer Vision, with the ultimate goal of building multimodal AI systems that empower users through highly contextualized assistance. This requires a leap from passive observation to active partnership, beginning with event-aware perception to transform raw video into a structured understanding of human actions.

By bridging the gap between recognition and reasoning, my work seeks to infer human goals and intentions, allowing AI to move beyond simple labeling and toward anticipating a user’s needs in real-time. These capabilities are fundamental to the development of embodied agents capable of navigating dynamic environments and performing meaningful, supportive tasks. We pursue scaling these multimodal technologies to solve high-impact, practical problems that enhance how humans and machines interact.

news

Feb 2026 I am a Lead Area Chair for ECCV 2026.
Nov 2025 I am a Lead Area Chair for CVPR 2026.
Oct 2025 I am an invited speaker and panelist at the ICCV 2025 Workshop on Multi-Modal Reasoning for Agentic Intelligence. Check out my slides here.

latest posts

selected publications

  1. strefer_ICCVW2025.png
    Strefer: Empowering Video LLMs with Space-Time Referring and Reasoning via Synthetic Instruction Data
    Honglu Zhou, Xiangyu Peng, Shrikant Kendre, Michael S Ryoo, Silvio Savarese, Caiming Xiong, and Juan Carlos Niebles
    In ICCV Workshop on What is Next in Multimodal Foundation Models?. Honolulu, Hawaii. Oct 2025
  2. lamsim-acl2025.png
    LAM Simulator: Advancing Data Generation for Large Action Model Training via Online Exploration and Trajectory Feedback
    Thai Quoc Hoang , Kung-Hsiang Huang, Shirley Kokane, Jianguo Zhang, Zuxin Liu, Ming Zhu, Jake Grigsby, Tian Lan, Michael S Ryoo, Chien-Sheng Wu, and 5 more authors
    In ACL Findings. Vienna, Austria. Jul 2025
  3. viunit_cvpr25.jpg
    ViUniT: Visual Unit Tests for More Robust Visual Programming
    Artemis PanagopoulouHonglu Zhou, Silvio Savarese, Caiming Xiong, Chris Callison-Burch, Mark Yatskar, and Juan Carlos Niebles
    In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Nashville, Tennessee. Jun 2025

all publications