Emilien Seiler

Curriculum Vitae

Emilien Seiler · PhD Student in Computer Vision, Computer Vision Laboratory (CVLab), EPFL

Download CV (PDF)

Education

2025 – present
Ph.D. in Computer and Communication Sciences (EDIC), EPFL, Computer Vision Laboratory · Lausanne, Switzerland
  • Advisor: Prof. Pascal Fua. Supported by the EDIC Fellowship.
  • First author of Flow-Corrected Shape Optimization (NeurIPS 2026), a state-of-the-art framework for shape optimization in high-dimensional 3D generative models that leverages flow matching.
2023 – 2024
Master's Thesis, Harvard University, IACS · Boston, USA
2021 – 2024
M.Sc. in Computational Science and Engineering, EPFL · Lausanne, Switzerland
2018 – 2021
B.Sc. in Environmental Sciences and Engineering, EPFL · Lausanne, Switzerland

Experience

2025
Data Engineering Intern, Liom · Zürich, Switzerland
  • Co-architected and deployed an end-to-end cloud data platform automating data pipelines for non‑invasive glucose monitoring R&D.
2022 – 2023
Data Science Intern, Jules-Gonin Eye Hospital · Lausanne, Switzerland
  • Engineered deep learning models on high-resolution OCT scans to discover 9 clinical biomarkers of CSCR recurrence; published as first author in Scientific Reports (2024).
  • Co-developed Cohort Builder, a software pipeline that assembles patient cohorts from medical records and ophthalmic imaging data (MIE 2024).
2021 – 2022
Research Assistant, EPFL · Lausanne, Switzerland
  • NLP Lab: Trained T5 transformers via actor-critic reinforcement learning.
  • CVLab & SPC: Predicted nuclear fusion disruptions with deep learning.

Teaching

2025 – present
Research Project Supervisor, EPFL
  • Supervised several students on research semester projects at the Computer Vision Laboratory.
2019 – 2023
Teaching Assistant, EPFL
  • Mathematical Analysis, Computer Science, Probability and Statistics, Mechanical Physics.

Publications

See the full list of publications →

Skills

Research
3D generative models · Flow matching & diffusion · Latent-space optimization · Shape representation (SDF, meshes, point clouds) · Differentiable rendering · Physics-informed ML
ML & Training
PyTorch · Hugging Face · Multi-GPU training (DDP) · SLURM / HPC clusters
3D
PyTorch3D · nvdiffrast · trimesh · Open3D · Blender (Python)
Engineering
Python · C++ · CUDA · Linux · Git · Docker · GCP
Languages
French (native) · English (fluent, C1)