Rosen Ting-Ying Yu

PhD student in Computational Science & Engineering (CSE) at MIT.

Rosen_v2.jpeg

Cambridge, MA

rosenyu [at] mit [dot] edu

I’m currently a PhD student in Computational Science & Engineering (CSE) at MIT. I’m a member of the Design Computation and Digital Engineering (DeCoDE) Lab, advised by Professor Faez Ahmed. My recent research focus is on developing Tabular Foundation Models (TFMs) and ML methods that build upon it (see this Nature article to know more about TFMs). I’ve published works on TFM-based Bayesian optimization (AutoML algorithms), multi-fidelity regression, fine-tuning with data generation pipelines. I also spent time at Prior Labs in Germany as their first research scientist intern (mentored by Professor Frank Hutter), contributing to the TabPFN ecosystem.

Prior to MIT, I received my B.S. in Electrical Engineering with minors in Computer Science (AI) and Geophysics from the Georgia Institute of Technology. I briefly worked on (geo)physics simulation research during my undergraduate time.

research highlights

May 20, 2026 Awarded the IMC PhD Fellowship from IMC Trading.
Apr 30, 2026 FIRE accepted to ICML 2026 (Spotlight, top 2.2%)! 🔥
Apr 28, 2026 Awarded the MIT CCSE Mathworks Fellowship.
Apr 01, 2026 FIRE accepted to the ICLR 2026 Workshop FM4Science (Poster).
Jan 25, 2026 GIT-BO accepted to ICLR 2026 (Poster)!

highlighted publications

  1. TFM
    ICML 2026 Spotlight
    2026_Yu_FIRE.png
    FIRE: Multi-fidelity Regression with Distribution-conditioned In-context Learning using Tabular Foundation Models
    Rosen Ting-Ying Yu, Nicholas Sung, and Faez Ahmed
    In ICML 2026 Main Conference (Spotlight), 2026
  2. TFM & BO
    ICLR 2026
    2025_yu_gitbo.png
    GIT-BO: High-Dimensional Bayesian Optimization with Tabular Foundation Models
    Rosen Ting-Ying Yu, Cyril Picard, and Faez Ahmed
    In ICLR 2026 Main Conference, 2026
  3. TFM
    EurIPS 2025
    2025_TabPFN2.5.png
    TabPFN-2.5: a Preview
    Leo Grinsztajn, Klemens Flöge, Oscar Key, and 12 more authors
    In EurIPS Workshop on AI for Tabular Data, 2025