AgentCollusionAI security startupCareers
Research this problem with us.
Join an AI security startup investigating hidden coordination and collective behavior in systems of interacting AI agents.
The problem we are hiring around.
Agent networks are moving from experiments into infrastructure. The hardest failures will emerge across relationships, incentives, memory, tools, and time.
This is a chance to help define the concepts, evaluations, and open questions for a problem that does not yet have a standard playbook.
We are looking for people who enjoy moving between theory, experiments, and systems thinking, and who can make uncertain problems concrete without pretending the answers are settled.
What you will work on
- Adversarial evaluation research
Design environments where strategic agent coordination can emerge.
- Relationship-level analysis
Study messages, tools, memory, incentives, and outcomes as one connected system.
- Governance questions
Explore what explainable, proportionate responses to suspected collusion could look like.
Open roles
Founding Researcher / EngineerResearch and engineering, open globallyTo apply, send a short note about a hard technical problem you pursued, what you built or learned, and why multi-agent security matters to you. Links to code, papers, or experiments are welcome.
