Staff Security Engineer - AI
Overview:
In this role you will own the security of Compass's AI-enabled products, shaping the multi-year strategy to secure AI capabilities across the platform. You will collaborate with AI platform and product teams to unify security standards, perform threat modeling, and deploy real-time protections for AI inputs/outputs. Your work will reduce risk in RAG pipelines, agent workflows, and retrieval systems, while guiding executives on AI risk and engineering roadmaps. This role combines security, AI, and platform-scale impact in a fast-growing tech organization. Compensation / Benefitsmedical, dental and vision benefits 401(k) planpaid vacation and holidaysparenteral leave and recharge leaveflexible spending accounts (FSAs)equity or incentive programs
Responsibilities:
- Own the AI product security roadmap and define vision and standards with AI platform and product engineering teams
- Threat model AI capability builds, conduct architectural reviews for AI features, RAG pipelines, multi-agent workflows, and automated decisioning
- Build production guardrails for LLM inputs/outputs: injection mitigation, jailbreak defense, output verification and data egress checks
- Design blast-radius controls for autonomous agents: permissions, human-in-the-loop, audit trails, safe tool invocation
- Secure the retrieval layer: access controls, source integrity, tenancy boundaries, and retrieval protections across RAG/vector stores
- Harden the AI supply chain: assess posture, privacy controls, data retention, and terms with third-party models/providers
- Instrument and detect AI abuse in production: telemetry for tool invocation, prompt abuse, data exfiltration; integrate with monitoring
- Automate secure AI patterns: reusable libraries, CI/CD checks, evaluation harnesses
- Advise and evangelize: translate AI risk into engineering roadmaps for engineers and executives
Key requirements:
- 7+ years in security engineering with 2+ years in AI/ML or LLM security
- End-to-end security of production AI/ML systems including lifecycle and runtime
- Experience with adversarial ML and LLM risks; hands-on red teaming
- Strong application and cloud security fundamentals (AWS and/or GCP, Kubernetes, IAM, API security, DevSecOps)
- Experience threat modeling complex distributed systems and remediating with engineering teams
- Strong strategic communication skills translating AI risk into business impact
- Proactive and solution-oriented mindset
- Clear cross-functional communication
- Capability to translate complex AI risk into actionable roadmaps
- Python or Go production coding
- Experience building security pipelines and integrations for AI/ML workloads
- Threat modeling and secure-by-default AI patterns