ServiceNow AI Research

Tomás Vergara Browne

Tomás Vergara Browne

Visiting Researcher

Agentic Harness & Defenses

I am a PhD student at Mila and McGill, working with Siva Reddy and Marius Mosbach. I am interested in (1) rigorously understanding the capabilities and information inside machine learning models, and (2) building models that maximize reasoning capabilities, without relying on memorization (a cognitive core).

Interests
  • Theory of Machine Learning
  • Reasoning
  • Natural Language Processing

Publications

Koala Science: A Platform for AI Reviewers in the Wild. Tomás Vergara Browne, Ada Tur, Jay Gala, Mehar Bhatia, ziling cheng, Fengyuan Liu, Gaurav Sahu, gaurav kamath, Zihan Wang, Amirhossein Kazemnejad, Arkil Patel, Charbel El Feghali, Emiliano Penaloza, erfan darian, Kusha Sareen, Martin Weiss, Marzia Nouri, Parishad BehnamGhader, Shubham Gupta, Weitze Suijker, Marco Ciccone, Aditi Khandelwal, Xing Han Lu, Zichao Li, Cesare Spinoso-Di Piano, Darshan Patil, David Dobre, David Heurtel-Depeiges, Johan Obando, Mandana Samiei, Lu Xiusi, Stephane Gazaille, Saurav Jha, Baptiste Bonin, Elham Daneshmand, Ya-Shi Zhang, Jianan Zhao, Xixian Liu, Laurent Charlin, Christopher Pal, Hugo Larochelle, Benno Krojer, Spandana Gella, Siva Reddy. At Workshop at the Neural Information Processing Systems (NeurIPS),  2026.

Cite