Tools for AI auditors and
AI safety/security researchers.
Our mission is to give auditors, both internal and third-party, and AI safety/security researchers superior tools to understand AI agent activity, identify harmful or misaligned behavior, and assess whether control measures are good enough to prevent harm. Our tools are open by default, designed to run in the user's own environment, and work for time-constrained audits.
Working with
on Control Tower
used by
,
which maintains a public fork of our proxy
Our work addresses two problems: investigation and testing.
Investigation
It is difficult to reliably work out what agents did. The evidence is too large, split across coordinating swarms of agents, sometimes deliberately obscured, and possibly incomplete or altered.
Tools to analyze agent activity, verify log integrity, and identify harmful behavior.
Testing
Control evaluations don't reflect how agents are deployed inside the labs. They run outside a realistic deployment/harness environment, and sabotage examples are either too obvious or too harmless. As a result, controls appear stronger than they actually are.
Tools to red-team AI lab systems and controls in realistic deployment environments, allowing auditors to assess whether safety controls work.
We will develop these capabilities with direct feedback from auditors (e.g., Faculty) and AI control researchers (e.g., Redwood Research). Over time, we aim to improve agent incident detection and measure how well controls perform across labs and models.
Co-founded Luthien in early 2025, built an open-source proxy, and raised $125k in investment. Before Luthien, he spent 9 years at Amazon, mostly building AI products as a principal technical product manager, and drove $4.5B in profit.2
Co-founded Lamar Health (Y Combinator) as CTO and led machine learning, engineering and ops teams at McKesson. He spent 13 years as a software and ML engineer after 12 years in product management at Adobe and Autodesk.
Cyber threat intelligence leader with public-sector and intelligence-community experience. Has built AI threat models and evaluations for AI safety research organizations.
Built benchmark quality-assurance tooling at MARS and risk-modeling methods with SaferAI at SPAR. He spent 5 years as a biostatistician at the University of Pittsburgh and holds an MA in Statistics.
1 Conditional on Luthien funding and full-time offer.
2 All figures are based on A/B experiment results, annualized, and include long-term sales/profit based on synthetic controls ($1.6B) and NPV of future cash flows at a 10% discount rate ($3.5B).
We're hiring a founding Member of Technical Staff. If you want to build tools that help auditors and AI safety researchers keep up with AI agents, we'd like to hear from you.
See the open role