About
PhD researcher in machine learning/optimization. Interested in quasi-Newton and second‑order optimization methods, and efficient/low‑precision training.
On top of that, I'm a big fan of tennis and badminton.
Skills
- Python, PyTorch, NumPy, Pandas, SciPy, Matlab
- C++ (working)
- Optimization: first/second‑order
- Quantization / low‑precision training
- Experiment design & analysis
- Linux, HPC/SLURM, basic profiling
Education
University of Basel — PhD, Machine Learning / Optimization
Advisor: Aurelien Lucchi.
Technical University of Munich (TUM) — M.Sc., Electrical Engineering
Top 1%; GPA 5.93/6.0; focus on biomedical imaging & signal processing.
Technical University of Munich (TUM) — B.Sc., Electrical Engineering
Top 1%; GPA 5.93/6.0
Selected Publications
- Theoretical characterization of Gauss-Newton conditioning in Neural Networks. NeurIPS, 2024. pdf · code Zhao, J.*, Singh, SP.*, Lucchi, A.
- Cubic regularized subspace Newton for non-convex optimization. AISTATS, 2025 (Oral). pdf · code Zhao, J.*, Lucchi A.*, Doikov, N.*
- Deep optoacoustic localization microangiography of ischemic stroke in mice. Nature Communications 14 (1), 3584 pdf · Deán-Ben, X., Robin, J., Nozdriukhin, D., Ni, R., Zhao, J., Glück, C., Droux, J., Sendón-Lago, J., Chen, Z., Zhou, Q., Weber, B., Wegener, S., Vidal, A., Arand, M., El Amki, M., Razansky, D. *equal contribution on items 1-2
Contact
Email: jim.zhao at unibas.ch
For opportunities: please include role, team, location, and start window.