Kevin Scaman

Research scientist in machine learning, optimization and graphs at Inria and ENS Paris

Profile photo

I am a research scientist at INRIA Paris and the DI ENS (computer science department of Ecole Normale Supérieure de Paris) in the Argo team, and a part-time associate professor (chargé de cours) at Ecole Polytechnique. My research lies at the intersection of machine learning, optimization and graph theory. Prior to that, I was a research scientist in Huawei’s machine learning lab in Paris leading the optimization for ML project-team. I did a post-doc in distributed optimization at the Microsoft Research – Inria Joint Center, and obtained my PhD in Applied Mathematics from the Ecole Normale Supérieure Paris-Saclay under the supervision of Nicolas Vayatis. In 2018, I received a best paper award of the 2018 NeurIPS conference for my work on distributed optimization.

CV contact LinkedIn Google Scholar Twitter

Education

  • 2013 - 2016

    PhD in Machine Learning
    Ecole Normale Supérieure Paris-Saclay

  • 2011 - 2012

    Msc in Machine Learning (MVA)
    Ecole Polytechnique

  • 2008 - 2011

    Eng. deg in Applied Mathematics
    Ecole Polytechnique

Experience

  • 2022 - now

    Part-time associate professor
    Ecole Polytechnique, Saclay

  • 2021 - now

    Researcher
    INRIA and ENS, Paris

  • 2018 - 2021

    Principal researcher
    Huawei Noah's Ark, Paris

  • 2017 - 2018

    Post-doctoral researcher
    Microsoft Research - INRIA joint center, Saclay

Awards

  • 2018

    Best paper award (4 out of 4865 submissions)
    NeurIPS 2018, Montréal

  • 2020

    Individual gold medal award
    Huawei, Paris

  • 2019

    Outstanding contributions individual award
    Huawei, Paris

  • 2018

    Future star award
    Huawei, Paris

Research

My field of study is at the intersection of Machine Learning, structured data analysis (graphs, time series) and optimization. A common theme of my research consists in analyzing the impact of structure in data (spatial proximity, time dependency or item correlations) and using it for ML purposes (training large-scale machine learning models, improving low-dimensional representations of graphs, identifying patterns from malware traces, or predicting the outcome of information cascades).

Teaching

  • 2023-2024

    Mathematics of Deep Learning [course material]
    M2 MASH, PSL, Paris

  • 2023-2024

    Deep Learning [course material]
    ENSAE, Saclay

  • 2023-2024

    Deep Learning
    ENS, Paris

  • 2023-2024

    Deep Learning
    Ecole Polytechnique, Saclay

  • 2022-2023

    Machine Learning 2
    Master X-HEC, Ecole Polytechnique, Saclay

  • 2022-2023

    Capstone projects
    M2 Data Science, Ecole Polytechnique, Saclay

  • 2022-2023

    Mathematics of Deep Learning [course material]
    M2 MASH, PSL, Paris

  • 2022-2023

    Deep Learning
    ENS, Paris

  • 2021-2022

    Deep Learning (mathematics track)
    ENS, Paris

  • 2021-2022

    Structures et Algorithmes Aléatoires (TDs)
    ENS, Paris

Students supervision

I have had the pleasure to supervize the work of amazing students and junior researchers.

Publications

My publications are also listed on Google Scholar and DBLP.

Contact

Email: kevin.scaman@inria.fr

Office: INRIA Paris, 2 rue Simone IFF, 75012 Paris