Jike Zhong

I am a second-year PhD student at the University of Southern California (USC). I work primarily with Prof. Konstantinos Psounis (he is fantastic!) on multimodal foundation models, I also work closely with Prof. Sai Praneeth Karimireddy. My recent research focuses on visual and multimodal reasoning, post-training with reinforcement learning, and post-training for personalized intelligence.

Before PhD I spent a year with the data science team at Salesforce working on LLM products. Prior to that I completed Bachelor's in CS and Stat at Ohio State University (OSU, 2019-2022), advised by great Prof. Wei-Lun Chao and Dr. Hong-You Chen. I have also spent some wonderful time at CCVL lab at Johns Hopkins University (JHU), supervised by Bloomberg Distinguished Prof. Alan Yuille.

I'm deeply drawn to research that enhances our fundamental understanding of large models, from data curation, pre-training, to post-training. We live in an exciting time of AI revolution, if you have ideas to discuss or collaborate, feel free to reach out! I can be reached at jikezhon@usc.edu

Last update was in 09/2025.

Email  /  Google Scholar

• 2025-09: To Think Or Not Think accepted to NeurIPS 2025 as Spotlight
• 2025-07: New paper on autoregressive motion prediction will appear as Oral in MICCAI '25 workshops
• 2025-05: Internship with congnition team at HRI with Prof. Shao-Yuan Lo
• 2025-04: We release To Think Or Not Think and show that thinking is not always beneficial for RL post-training
• 2025-04: We release Med-R1, reinforcement finetuning for medical MLLM reasoning
• 2025-03: New paper on benchmarking multimodal performance on EEE accepted at CVPR '25
• 2024-09: New paper on vision-language finetuning accepted at EMNLP '24
• 2024-08: Started PhD at USC