AI Security & Control Researcher
Apollo Research
AI Security & Control Researcher
Apollo Research
Audits AI models for deceptive alignment
THE OPPORTUNITY
Apollo Research works with most frontier AI companies (OpenAI, Anthropic, Google, Meta, Thinking Machines and others) to test their models before deployment and collaborate on fundamental scheming research. Our coding agent security product, Watcher, is deployed in production and monitors billions of agent tokens per month across engineering teams at agent-building scale-ups and enterprises.
We are looking for a security & control expert to help us design better threat models and control protocols against AI adversaries, and improve the effectiveness and security of Watcher.
This is truly a "start-up role" in the sense that you have significant say in shaping the direction of the role. This is an individual contributor role but could lead to management responsibilities eventually, if desired.
Develop threat models for coding agents under the assumption that they could be misaligned or compromised. This includes near-term threats like AI agents leaking private internal data and long-term threats like agents sabotaging safety research.
Treat coding agents through the lens of insider risk: they have credentials, access to code, network access, and the ability to execute arbitrary actions, just like a malicious insider would (see Control agenda).
Map out kill chains and attack progressions similar to frameworks like MITRE ATT&CK, adapted for agentic AI. See e.g. the Agentic Loss-of-Control Threat Matrix for an example of a high-quality contribution.
Build and maintain our “coding agent security levels” which define what level of robustness Watcher provides against different categories of failure modes. These levels should be concrete, testable, and usable both internally (to guide product priorities) and externally (to communicate our security posture to customers).
Maintain our library of coding agent failure modes and keep it comprehensive, accurate, and current.
Stay up to date with the academic and industry literature on coding agent security, AI security, and adjacent fields (prompt injection, deception, accidental security implications, supply chain attacks on AI systems, etc.).
Design realistic attack trajectories and example exploits for monitor development and backtesting, work with the research team to ensure monitors are trained and evaluated against credible adversarial behavior.
Serve as security adjudicator for flagged trajectories. When monitors flag ambiguous agent behavior, determine whether it represents a genuine attack, a benign-but-unusual workflow, or needs deeper investigation, and feed those judgments back into monitor improvement.
Red-team and improve Watcher's monitors and policies, document findings, and feed them into product improvement.
Work with the technical research staff to build adversarial test suites into the backtesting pipeline.
*This role is distinct from the AI Security Researcher role on Apollo's Infra & Security team, which threat-models Apollo's own internal systems and the agents operating on them.
Develop a comprehensive coding agent threat model: Think deeply about all the ways in which coding agents could attack an organization. Compare this threat model to conventional threats from human insiders. Publish a detailed research piece describing the threat model building on existing research, e.g. from Redwood Research.
Improve our database of failure modes: We have an internal database of 50+ failure modes of coding agents with detailed reports for all of them. For this project, you would provide an expert view on the current state of that database and suggest improvements. In the long run, you would maintain that database and be responsible for integration of new failures.
Prioritize failure modes that Watcher should cover: Different parts of Watcher attempt to cover different threat models and attack strategies. Based on the results of the threat model project above, we want to ensure that each part of Watcher covers the most important failure modes in the most efficient way. For example, not all monitors require blocking capability and some failure modes might benefit from additional affordances like being able to disperse subagents.
Experience with AI/ML systems security, LLM security, or AI control research. The field is young enough that deep experience here is rare, but any exposure significantly reduces ramp-up time.
Detection engineering, SOC, or incident analysis experience. A part of this role is judging whether flagged agent behavior is genuinely malicious, and people who have triaged real-world alerts might ramp much faster.
Familiarity with insider threat programs or insider risk frameworks. The mental model of "the coding agent is a potentially malicious insider" is useful for this role and someone who has worked on insider threats will pick it up faster.
Red teaming or offensive security background. Useful for the Watcher red-teaming responsibilities and for thinking adversarially about failure modes.
Formal AI safety research background. Helpful but not necessary. We need security practitioners who can learn the AI safety context, not AI safety researchers who need to learn security.
Management experience. This is an IC role, at least initially.
Specific certifications (CISSP, etc.). We care about demonstrated ability, not credentials.
The product team consists of research scientists: Victor Gillioz, Monika Jotautaitė, Dmitrii Volkov; product engineers: Jeremy Neiman, Zak Walters, Zen van Riel, Srdjan Miletic and Gustavo Bicalho; and our GTM lead: Kyle Dai. Marius Hobbhahn (CEO) advises the team. Furthermore you will interact with our other SWEs and researchers, since we intend to be "our own customer" by using our products internally for our research work. You can find our full team here.
The rapid rise in AI capabilities offers tremendous opportunities, but also presents significant risks. At Apollo Research, we're primarily concerned with risks from Loss of Control, i.e. risks coming from the model itself rather than e.g. humans misusing the AI. We're particularly concerned with deceptive alignment / scheming, a phenomenon where a model appears to be aligned but is, in fact, misaligned and capable of evading human oversight.
We work on the science of scheming, detection of scheming (e.g. building evaluations), and scheming mitigations (e.g. anti-scheming). We also work on control and monitoring research (see our scalable monitoring agenda). We work closely with many frontier AI companies, such as OpenAI, Anthropic, Google, Meta, Thinking Machines and others, e.g. to test their models and collaborate on the science of scheming. At Apollo, we aim for a culture that emphasizes truth-seeking, being goal-oriented, giving and receiving constructive feedback, and being friendly and helpful. If you're interested in more details about what it's like working at Apollo, you can find more information here.
We also build a coding agent security product called Watcher that secures agent deployments in companies. Our goal is to reduce the probability of catastrophic incidents by securing coding agents, learning about their real-world risks, and publishing our research on how to build these control systems most effectively.
Equality Statement: Apollo Research is an Equal Opportunity Employer. We value diversity and are committed to providing equal opportunities to all, regardless of age, disability, gender reassignment, marriage and civil partnership, pregnancy and maternity, race, religion or belief, sex, or sexual orientation.
Please complete the application form with your CV. The provision of a cover letter is neither required nor encouraged. Please also feel free to share links to relevant work samples.
About the interview process: Our multi-stage process includes a screening interview, a take-home test (3 hours), 3 technical interviews, and a final interview with Marius (CEO). There are no leetcode-style general coding interviews. You may use AI tools on the take-home; we judge the result the way we'd judge any contributor's work, so you are responsible for the quality of everything you submit. If you want to prepare, we suggest building simple monitors for coding agents and running them on your own Claude Code / Cursor / Codex / etc. traffic.
Your Privacy and Fairness in Our Recruitment Process: We are committed to protecting your data, ensuring fairness, and adhering to workplace fairness principles in our recruitment process. To enhance hiring efficiency, we use AI-powered tools to assist with tasks such as resume screening. These tools are designed and deployed in compliance with internationally recognized AI governance frameworks. Your personal data is handled securely and transparently. All resumes are screened by a human and final hiring decisions are made by our team. If you have questions about how your data is processed or wish to report concerns about fairness, please contact us at [email protected].
Audits AI models for deceptive alignment
Apollo Research is an AI safety organization that audits high-risk failure modes in large AI models, with a focus on deceptive alignment where models pretend to be aligned while pursuing their own objectives. It conducts fundamental research in interpretability and behavioral model evaluations, then applies these insights to audit real-world models. By combining examination of model internals with behavioral assessments, Apollo Research aims to provide stronger safety assurances than relying on behavioral evaluations alone. Its goal is to minimize catastrophic risks associated with advanced AI systems and ensure more reliable, safe behavior in deployed models.
Explore more opportunities at Apollo Research