Jinkun Cao
San Francisco, California, United States
Summary
🤩
RockstarJinkun Cao is a robotics PhD and research-focused engineer with nine years of experience building vision-first and physics-aware systems for human motion, interaction, and loco-manipulation. He combines academic rigor from Carnegie Mellon with industry research stints at Meta, NVIDIA, Adobe, and Amazon Frontier AI & Robotics, translating novel models into practical tracking, animation, and control tools. His work spans learning-based and simulation-informed approaches, from RL-driven motion animation to hand-object interaction and robust multi-object tracking contributions (notably to the OC-SORT project for occlusion-robust MOT). Based in San Francisco, he ships both research code and polished engineering—often bridging dataset evaluation, algorithm tweaks, and deployment details. Outside core research he publishes code and photographs publicly, reflecting a habit of documenting technical and creative work.
10 years of coding experience
Stackoverflow
233reputation
409kreached
4answers
2questions
Github Skills (17)
python10
deeplearning-ai10
deep-learning10
object-detection10
computer-vision10
numpy9
opencv8
pytorch8
decision-tree6
calculation6
neural-network6
regression6
anaconda6
tensorflow6
machine-learning6
Programming languages (7)
Github contributions (5)
noahcao/OC_SORT
Mar 2022 - Nov 2022
[CVPR2023] The official repo for OC-SORT: Observation-Centric SORT on video Multi-Object Tracking. OC-SORT is simple, online and robust to occlusion/non-linear motion.
Role in this project:
ML Engineer
ML EngineerContributions:25 commits, 4 PRs, 36 pushes in 7 months
Contributions summary:Jinkun primarily contributed to the object tracking functionality of the OC-SORT project. Their commits involve modifications to core tracking algorithms, including the OCSort class, and integration with public detection results. The user also worked on evaluation scripts, particularly for the DanceTrack dataset, and introduced features like the use of byte tracking. These changes suggest a focus on improving the performance and adapting the tracking system.
multi-object-trackingocclusiontrackingcomputer-visiondeep-learning
SysCV/spc2
Jan 2021 - Dec 2022
Instance-Aware Predictive Navigation in Multi-Agent Environments, ICRA 2021
Contributions:12 commits, 6 PRs, 5 pushes in 1 year 11 months
multi-agent
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.