RLEval: Methods and Reinforcement Learning Environments for Evaluating AI Agents.

Jonas Mueller

Handshake

Leads a group advancing Data and Evaluations for frontier AI. Previously at Cleanlab, where he pioneered methods to detect and remediate issues in enterprise AI agents. Earlier, he was a Senior Scientist at AWS building AutoML and deep learning services. He completed his PhD at MIT CSAIL and undergraduate studies at UC Berkeley.

Anish Athalye

Handshake

Leads research on RL environments and post-training. Previously co-founder and CTO at Cleanlab (acquired by Handshake). He completed his PhD at MIT CSAIL in the PDOS group, advised by Frans Kaashoek and Nickolai Zeldovich.

Priyaranjan Pattnayak

Oracle AI

Senior Principal Scientist at Oracle Cloud, leading research initiatives in Generative AI, NLP, and AI robustness for enterprise-scale multilingual and multimodal systems. His work emphasizes production-grade agentic frameworks, low-resource language modeling, evaluation, and trustworthy AI deployment.

Aziza Mirsaidova

Oracle AI

Works in GenAI evaluations across multimodal, text, and code generation. Also an Ethics and Technology Practitioner 2026 Fellow at Stanford. Previously worked in AI safety at Microsoft Responsible AI research.

Natasha Jaques

University of Washington, Google DeepMind

Focuses on social reinforcement learning in multi-agent and human-AI interactions. During her PhD at MIT, she developed foundational RLHF techniques for language models. Her work has received awards at NeurIPS and ICML and has been widely featured in research and media venues.

Andi Peng

humans&

Previously worked on RL and post-training at Anthropic across Claude 3.5 to 4.5, and completed her PhD at MIT CSAIL advised by Jacob Andreas and Julie Shah. Her work centers on agents that learn representations from rich human knowledge.

Rasool Fakoor

Boson AI

Leads reinforcement learning and modeling efforts for robust agentic systems across changing tasks and environments. Previously a researcher and tech lead at AWS, and a founding member of post-training efforts for Amazon’s early RLHF models.

Ahmed Elgohary

Microsoft Research

Ahmed is Senior Researcher at Microsoft Research AI Frontiers focusing on reasoning and agentic models. Until very recently, he was part of the Microsoft Responsible AI team where he worked on the safety alignment and evaluation of LLMs and Agents.

Alina Gavrilov

Stanford University

Alina is a graduate student in Computational and Mathematical Engineering at Stanford University, researching women's health modeling using wearables. Her previous experience is in structural civil engineering and project management in local government. She is interested in designing AI systems for public good that are measurable, interpretable, and robust to uncertainty.

Holds a PhD in natural language processing from the University of Melbourne. Her work focuses on evaluation and reliability of machine learning and large language models, including human-centered and uncertainty-aware approaches.

Miguel Ballesteros

Oracle AI

Miguel is currently an AI Architect (Applied Scientist) at Oracle Cloud Infrastructure (OCI). Previously, he worked as a Principal Applied Scientist at AWS, and has also held positions at IBM, Carnegie Mellon University (CMU), and Universitat Pompeu Fabra (UPF). He has taught undergraduate and graduate courses at both CMU and UPF.

Graham Horwood

Oracle AI

Graham is Senior Director of Applied Science at Oracle’s OCI AI Science team, leading applied science efforts at the intersection of large-scale ML systems and enterprise cloud AI. Prior to Oracle, he served as Senior Director of Applied Science at Amazon Web Services (AWS).