Alexander Pondaven
🎓 PhD student at Oxford’s Torr Vision Group working on controllable video generation - keen on making generations more interactive (or even playable)
🔬 Research
- PhD (2023-Present): Working on controlling video diffusion models at Oxford AIMS CDT, generously funded by Snap studentship
- MEng Project (2022-2023): Enhanced Stable Diffusion 2 sampling diversity using particle-based methods (project · slides)
- Climate AI (2022): Developed meta-learning framework for satellite image inpainting, published at NeurIPS 2022 climate workshop (paper · project)
💼 Industry Experience
- Humanising Autonomy (6 month internship in 2022): Built active learning system and AWS data pipeline for video object detection training
- MathWorks (3 month internship in 2021): Developed video action recognition demo and enhanced pooling layers for spatio-temporal operations in 2022a release of MATLAB Deep Learning Toolbox