About this role
Summary: As the first AI security researcher on the product team, you will identify and mitigate risks in enterprise AI agents and models. You will conduct red teaming tests, build security guardrails, and publish research to inform product development and threat intelligence.
Alice secures AI apps, agents and models from build to production. Enterprises already run agents that talk to their customers, read their email and call their internal tools. Every conversation and every connection can be a way in for an attacker, or the point where an agent goes rogue. You find out how those agents break, on the stacks enterprises actually use, before they cause harm.
You will be the first AI security researcher on the product team, reporting to the Director of GenAI Products. Your research becomes guardrails and red teaming tests in the Alice platform, proof customers can act on, and research our marketing team publishes. It also feeds the threat intelligence we deliver. You choose the questions, guided by what enterprises are deploying right now and by the wider research Alice does with the foundation model labs.
Key Responsibilities
- Research emerging agent risks. Study how agents fail when they talk to users, read untrusted content, call tools, share memory, act autonomously and work with other agents. That covers indirect prompt injection, tool and MCP misuse, data exfiltration, memory poisoning, privilege escalation across agents and unsafe autonomous actions.
- Test and prove. Build the environments and experiments that show whether a risk is real, how often it happens and under which conditions. Work on the stacks enterprises use: agent frameworks, MCP servers, workflow tools and AI gateways. Build the sandboxes, detectors and graders that catch failures, and turn each one into a regression test anyone can rerun. Measure how often and how severely agents misbehave, and how that shifts across models, tools and harnesses. Report results with the numbers behind them, so others can rerun and trust them.
- Shape the platform. Work with the product team to turn your findings into new guardrails and red-team tests in the Alice platform.
- Inform go-to-market and threat intelligence. Share findings and evidence with sales, marketing and Alice's threat intelligence teams, so they can explain new risks to customers with proof behind them.
- Publish. Turn research into technical posts, reports and talks with the marketing team, written to hold up with security practitioners, buyers and analysts. We support conference submissions such as Black Hat and DEF CON, including time to prepare and travel to present. Vulnerabilities you find go through responsible disclosure and get CVEs where they qualify.
- Collaborate. Work with Alice's large research organization, which partners with leading hyperscalers and foundation model labs. Follow new papers, disclosures and exploits, and decide which ones matter for enterprises deploying agents.
Success Looks Like
90 days in, you have published at least one research piece on enterprise agent risks. Its findings have shaped new protections in the Alice platform.
- 4+ years in security research, offensive security, penetration testing or vulnerability research.
- Hands-on experience with direct and indirect prompt injection, jailbreaks, data exfiltration and tool misuse against real systems.
- Strong Python. You build your own tooling and test environments.
- Working knowledge of how agents are built: tool calling, MCP, retrieval, memory and multi-agent orchestration, with hands-on use of at least one agent framework.
- Familiarity with AI security frameworks (OWASP Top 10 for LLM Applications and for Agentic Applications, MITRE ATLAS) and with public agent safety and security benchmarks.
- Experiment discipline: reproducible setups, controls, and enough runs to trust a result.
- Clear writing and presentation: test results the product team can act on, and research that marketing can publish.
- Fluent English, written and spoken.
Nice to Have
- Published research, CVEs, conference talks, or a bug bounty or CTF record in AI or application security.
- Experience researching AI safety failures, such as harmful outputs or unsafe autonomous actions.
- A background in adversarial machine learning.
- Experience building or evaluating detection systems that use AI.
- Experience working with product and engineering teams to turn research into shipped features.
Alice is a trust, safety, and security company built for the AI era. We safeguard the communicative technologies people use to create, collaborate, and interact- whether with each other or with machines.
In a world where AI has fundamentally changed the nature of risk, Alice provides end-to-end coverage across the entire AI lifecycle. We support frontier model labs, enterprises, and UGC platforms with a comprehensive suite of solutions: from model hardening evaluations and pre-deployment red-teaming to runtime guardrails and ongoing drift detection.
Alice is widely considered a global leader in online safety and AI security. We have some of the most forward-thinking and passionate minds in the world working to safeguard over 3 billion users across the largest AI and tech platforms.
If you're creative and driven to secure the future of AI, we want to hear from you!