Teaching Assistance
I served as a teaching assistant or contributed to the teaching material for the following courses:
- Training Large Language Models — EPFL · Course description
- Mathematics of data: from theory to computation — EPFL · Course description
- Data Mining and Neural Networks — KU Leuven · Course description
- Support Vector Machines: Methods and Applications — KU Leuven · Course description
Mentoring
I (co-)supervised the following master's students:
- Adaptive Geometry-Aware Acceleration, by Hanka Goralija → EPFL PhD (AY 2025–2026, NeurIPS'26) · Paper.
- Dynamic Scaling and Spectral Preconditioning, by Matya Aydin → Google DeepMind (AY 2025–2026).
- Optimistic Online Learning for Data Mixing, by Sofija Orlovic (Spring 2026) · Paper.
- Distributed Orthogonalization in Muon, by Haocong Li (Fall 2026).
- Self-improving methods for LLM agents, by Son Le → Oracle (Spring 2025).
- Regularization Effects of LoRA in GRPO, by Antonin Rottman → UTokyo MS (Fall 2025) · Paper.
- Data mixing for LLM Knowledge Distillation, by Daniil Dzenhaliou → Tufa Labs (Fall 2025).
- Selective SSMs for graph signal processing, by Aidan Li → Mila MS (Summer 2024).
- Membership Inference Attacks against Vision-Language Models, by Zhan Li → Ant Group (Spring 2024, NeurIPS'24) · Paper.
- Multiway spectral clustering with kernel machines, by Edward Vandercruysse (AY 2021–2022).
- Non-local denoising of textured meshes, by Guillaume Roy (AY 2020–2021).