Member of Technical Staff - Research Scientist

About Patronus AI

Patronus AI is a frontier lab developing simulation research and infrastructure to accelerate progress toward human-aligned AGI. We are on a mission to simulate all of the world's intelligence. We are the team behind some of the earliest and most influential research in AI evaluation like Finance Bench, Lynx, Simple Safety Tests , Copyright Catcher, Humanity’s Last Exam, and more. Formerly AI researchers and engineers from companies like Meta AI, Amazon AGI, and Google.

Our customers include foundation model labs and Fortune 500 enterprises like Adobe. We are backed by top-tier investors including Lightspeed Venture Partners, Notable Capital, Stanford University, Noam Brown, Gokul Rajaram, and more.

As a Research Scientist at Patronus AI, you will own and drive foundational research that defines how agentic AI systems are trained, evaluated and improved. You will work at the intersection of reinforcement learning, simulations and scalable oversight, building systems that directly influence how frontier models are developed, stress‑tested, and deployed.

This is a highly autonomous role. You will tackle society’s most impactful research questions surrounding agent simulations, and translate them into rigorous experiments. You will have the opportunity to work on reward design, tool simulations and behavior analysis, shaping the industry standard for robust, high-quality environments. Your work will inform how frontier labs design, train and improve the next generation of agents for long-horizon tasks, progressing our path towards safe, human‑aligned general intelligence.

In this role, you will:

• Solve open-ended problems in scalable oversight, including agent cognition research, agent behavior analysis and new training methods for efficient RL.
• Develop state-of-the‑art systems for agent simulation, evaluation and reinforcement learning environments used to test and train frontier models.
• Drive novel research in RL and environment design, including reward design, trajectory evaluation, interruption handling, observation and state design, tool and action interface design, and curriculum learning.
• Publish papers, organize workshops and open source environments, datasets in collaboration with the evaluation research community.
• Independently scope, lead, and execute research projects end‑to‑end, from problem formulation and experiment design to results, write‑ups, and production impact.
• Design and run rigorous experiments, set timelines, and consistently deliver measurable research outcomes such as papers, benchmarks, datasets, and platform features.
• Keep up to date with the latest research literature and technologies in scalable oversight; synthesize findings into clear research updates and concrete system improvements.
• Experiment with new models, techniques, and tooling, proactively proposing and running experiments to improve our understanding of agent behaviors, learning signal quality, and RL scaling.
• Produce high‑quality, reproducible, production‑level research code that integrates cleanly into Patronus AI’s systems.
• Collaborate closely with research, engineering, and product teams to translate research ideas into robust, real‑world benchmarks and environments.
• Communicate progress clearly and proactively, flagging risks, blockers, or timeline changes early.
• Regularly review research progress with teammates, providing feedback, suggested next steps, and areas for improvement.
• Meet weekly with the CTO and researchers to align on priorities, progress, and next steps.
• Evangelize agent simulation and environment research internally and externally, contributing to Patronus AI’s thought leadership in scalable oversight.

• A BS, MS, or PhD in Computer Science, Machine Learning, Statistics, Mathematics, or a related quantitative field.
• Experience conducting independent research in reinforcement learning, NLP, agentic systems, evaluation, or alignment.
• Demonstrated ability to take open‑ended research problems from 0→1 and deliver high‑impact outcomes.
• Strong experimental skills, including experiment design, analysis, and interpretation of results.
• Experience writing clean,…