Hi there I am a Ph.D. student in CS at UB. My research focuses on machine learning, computer vision, robotics, and optimization. I also have a strong passion for contributing to open-source projects and creating impactful works in these fields.
- Email: [email protected]
- LinkedIn: zitong-zhan
- GitHub: zitongzhan
- PR #129856: a fast sparse linear system solver for CUDA GPUs, integrated into PyTorch's
linalglibrary. The optimization resulted in a 20-30% speedup over previous implementations.
- #184: Implemented batched EPnP (Efficient Perspective-n-Point) algorithm as a
torch.nn.Module - #216: Added ICP (Iterative Closest Point) module with batched input support
- #284: Enabled LieTensor operations to work under
func.vmap - #309: Preserved LieTensor type and operations under
vmap - #312: Added Conjugate Gradient (CG) solver for sparse linear systems
- #317: Implemented matrix multiplication between BSR and BSC sparse formats
- #320: Fixed
Inv()and other LieTensor operations error when usingpp.func.jacrev
- Conference: European Conference on Computer Vision (ECCV) 2024
- Code: iMatching GitHub
- Website: iMatching Website
- Journal: arXiv preprint arXiv:2409.12190 (2025)
- Code: bae GitHub
- Conference: IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) Workshop 2023
- Code: PyPose GitHub
- Video: Watch the demo
- Journal: arXiv preprint arXiv:2510.13310 (2025)
- Code: InstantSfM GitHub
- Website: InstantSfM Website
May 2025 - Aug 2025
Worked on Indoor SLAM for the Ring Always Home Drone.
- Production-grade loop closure module
- Lightweight feature aggregation and place recognition in home environments
- Robust indoor SLAM research for tricky textureless spaces
Feel free to reach out to me via email or connect with me on LinkedIn!