Yiyang Yu | 余易洋

Please find my full CV (PDF).

Experience

Senior Machine Learning Engineer — Meta, London, UK (2022 – present)

• Built and deployed ML systems for Trust & Safety, detecting policy violations at scale across messaging
• Drove end-to-end model development: feature engineering, training, offline/online evaluation, and production monitoring
• Led evaluation framework design to track model performance and behavioral drift over time

Ph.D. in Applied Mathematics — LPSM, Université Paris-Cité, Paris, France (2018 – 2022)

• Doctoral thesis: Deep Learning in Public Health, and Contributions to Machine Learning — covering statistical learning theory, random forest algorithms, and deep learning applied to healthcare prediction
• Supervised by Stéphane Gaïffas (LPSM, Université de Paris) and Emmanuel Bacry (CEREMADE, Université Paris-Dauphine)
• Co-developed WildWood, a random forest algorithm achieving gradient-boosting-level performance while remaining interpretable and lightweight — published in IEEE Transactions on Information Theory (code)
• Co-developed ZiMM, a deep learning model predicting long-term disease relapses from non-clinical claims data — published in Journal of Biomedical Informatics; presented at ML4H Workshop, NeurIPS 2019
• Funded by DIM Math Innov fellowship, Fondation Sciences Mathématiques de Paris
• Teaching Assistant at Université de Paris: Probability (18h) and Statistical Inference (42h) for third-year undergraduates

M2 Mathematics, Computer Vision, Machine Learning (MVA) — ENS Paris-Saclay, France (2017 – 2018)

Ingénieur Polytechnicien — École polytechnique, Palaiseau, France (2014 – 2018)