Feng Li
Shenzhen, Guangdong Province, China
Summary
🤩
RockstarFeng Li is a Senior Research Scientist and PhD candidate in Computer Science at HKUST with seven years of experience building state-of-the-art multimodal and dense vision models. He has driven research and engineering at top labs including Google DeepMind, Meta FAIR, Microsoft, and ByteDance, contributing to flagship projects like Segment Anything (SAM3) and open-source VLM efforts such as LLaVA variants and BAGEL. His hands-on work spans model development, post-training for Gemini-style multimodal systems, and production-ready code improvements—evident in notable open-source contributions to Mask DINO where he fixed core modules, dataloader issues, and checkpoints. Based in Shenzhen, he bridges rigorous research with practical engineering, routinely turning cutting-edge papers into robust implementations. Less obvious: Feng alternates between deep algorithmic advances and full-stack code hygiene, ensuring research prototypes are reproducible and deployable. He brings a rare mix of academic depth and production discipline to multimodal AI system building.
7 years of coding experience
Github Skills (6)
semantic-segmentation10
object-detection10
pytorch10
instance-segmentation10
python10
machine-learning9
Programming languages (2)
Github contributions (5)
IDEA-Research/MaskDINO
Jun 2022 - Jan 2023
[CVPR 2023] Official implementation of the paper "Mask DINO: Towards A Unified Transformer-based Framework for Object Detection and Segmentation"
Role in this project:
Full-stack Developer
Full-stack DeveloperContributions:1 release, 31 commits, 6 PRs in 7 months
Contributions summary:Feng's contributions primarily involve code releases, updates, and fixes within the Mask DINO project. They worked on updating the README, fixing duplicate modules, correcting the semantic data mapper, and updating the load checkpoint. This user also addressed issues related to license and dataloader functionalities.
pytorchmasksemantic-segmentationbackbonedino
IDEA-Research/awesome-detection-transformer
Mar 2022 - Dec 2022
Collect some papers about transformer for detection and segmentation. Awesome Detection Transformer for Computer Vision (CV)
Contributions:18 commits, 5 PRs, 17 pushes in 9 months
pytorchcollectvisiondetectiondeep-learning
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.