Ayoub Foussoul

Postdoctoral Researcher

Contact: ayoub.foussoul [at] chicagobooth [dot] edu
CV, Google Scholar, LinkedIn


Hello! I am a Postdoctoral Researcher at the University of Chicago Booth School of Business, working with Prof. Ozan Candogan. I am broadly interested in the intersection of deep learning and combinatorial optimization.

My recent work has focused on two directions:

  • The first is in predict-then-optimize settings where the downstream optimization problem is combinatorial. In these settings, standard deep learning architectures often lead to downstream optimization problems that are difficult to solve and hard to scale. My goal is to design architectures that are both expressive predictors and more compatible with downstream combinatorial optimization. In recent work, we introduced Deep Flow Networks, a new architecture for discrete predict-then-optimize problems that combines strong predictive performance with substantially faster downstream optimization than standard neural baselines.
  • The second is the use of deep learning to improve combinatorial optimization solvers. In ongoing work, we are developing a deep learning-based discrete local search algorithm that is cheaper and more amenable to parallelization than standard methods, while still producing high-quality solutions.
headshot

I received my Ph.D. from the Department of Industrial Engineering and Operations Research (IEOR) at Columbia University, where I was advised by Prof. Vineet Goyal. My doctoral work focused on developing approximation algorithms for sequential decision-making under uncertainty in supply chains and matching markets.

Prior to my Ph.D., I earned a Master’s and a Bachelor’s degree in applied mathematics from École Polytechnique in France.

I will be on the 2026–2027 academic job market.

Papers
  • Deep Flow Networks
    Ozan Cangocan & Ayoub Foussoul.
    Under review. (draft available on request).
  • Distributionally Robust Newsvendor on a Metric
    Ayoub Foussoul & Vineet Goyal.
    Major Revision in Operations Research.
    Extended abstract appeared in proceedings of EC 2025.
  • LP-based Approximations for Disjoint Bilinear and Two-Stage Adjustable Robust Optimization
    Omar El Housni, Ayoub Foussoul & Vineet Goyal.
    Mathematical Programming.
    Preliminary version appeared in proceedings of IPCO 2022.
  • Two-Stage Stochastic Stable Matching
    Ayoub Foussoul, Yuri Faenza & Chengyue He.
    In proceedings of IPCO 2024.
  • Minimum Cut Representability of Stable Matching Problems
    Ayoub Foussoul, Yuri Faenza & Chengyue He.
    Major Revision in Operations Research.
    Honorable Mention in INFORMS Optimization Society Student Paper Prize 2025
  • Fully-Dynamic Load Balancing
    Ayoub Foussoul, Vineet Goyal & Amit Kumar.
    Mathematical Programming.
    Preliminary version appeared in proceedings of IPCO 2024.
  • Last Switch Dependent Bandits with Monotone Payoff Functions
    Ayoub Foussoul, Vineet Goyal, Orestis Papadigenopoulos & Assaf Zeevi.
    In proceedings of ICML 2023.
  • MNL-Bandit in Non-Stationary Environments
    Ayoub Foussoul, Vineet Goyal & Varun Gupta.
Teaching
  • Optimization I, IEOR E6613 | Teaching Assistant & Guest Lecturer
    Core PhD course on linear and convex optimization. Fall 2021, 2022.
  • Optimization II, IEOR E6614 | Teaching Assistant
    Core PhD course on combinatorial optimization. Spring 2023, 2024, 2025.
  • Convex Optimization, IEOR E6616 | Teaching Assistant
    PhD course on convex optimization. Spring 2022.
  • Optimization Methods, IEOR E4004 | Teaching Assistant & Guest Lecturer
    Core Master’s course on optimization. Fall 2023.
  • Applications for Financial Engineering, IEOR E4500 | Teaching Assistant
    Master’s course on quantitative methods in financial engineering. Spring 2021.
Service
    Reviewer: Management Science, Mathematical Programming, Optimization Letters, NeurIPS MLxOR workshop, INFORMS Journal on Computing, IPCO, SODA, STOC, EC, ICML.
Talks
    Distributionally Robust Newsvendor on a Metric
    INFORMS Annual Meeting, October 2025, Atlanta GA
    Revenue Management and Pricing Conference (RMP), July 2025, New York NY
    ACM Conference on Economics and Computation (EC), July 2025, Stanford CA
    Manufacturing & Services Operation Management Conference (MSOM), June 2025, London
    Columbia University Data Science Day, April 2025, New York NY (poster session)
    IEOR Colloquium, November 2024, New York NY
    INFORMS Annual Meeting, October 2024, Seattle WA
    Cornell ORIE Young Researchers Workshop, October 2024, Ithaca NY
    Northwestern Kellogg Operations Management Rookiepalooza, October 2024, Evanston IL

    Two-Stage Stochastic Stable Matching
    Integer Programming and Combinatorial Optimization Conference (IPCO), July 2024, Wrocław
    Columbia University Data Science Day, April 2024, New York NY (poster session)

    Fully-Dynamic Load Balancing
    Integer Programming and Combinatorial Optimization Conference (IPCO), July 2024, Wrocław
    Columbia IEOR Student Seminar, February 2024, New York NY

    Last Switch Dependent Bandits with Monotone Payoff Functions
    International Conference on Machine Learning (ICML), July 2023, Honolulu HI (poster session)

    MNL-Bandit in Non-Stationary Environments
    DSI Financial and Business Analytics Poster Session, November 2023, New York NY (poster session)
    INFORMS Annual Meeting, October 2023, Phoenix AZ
    Revenue Management and Pricing Conference (RMP), July 2023, London

    LP-based Approximations for Disjoint Bilinear and Two-Stage Adjustable Robust Optimization
    International Symposium on Mathematical Programming (ISMP), July 2024, Montréal
    International Research and Innovation Seminar (IRIS), December 2023, Ben Guerir
    INFORMS Annual Meeting, October 2022, Indianapolis IN
    International Conference on Continuous Optimization (ICCOPT), July 2022, Bethlehem PA
    Integer Programming and Combinatorial Optimization Conference (IPCO), June 2022, Eindhoven
    INFORMS Optimization Society Conference (IOS), March 2022, Greenville SC