Replit
AI Agent Security Architect
Foster City, CA · Hybrid · Full-time · Posted Sep 1, 2026
- Experience
- 6+ years in security engineering or security architecture, with at least 2+ years specifically in AI/ML security, LLM application security, or securing agentic frameworks.
About the job
Designs and maintains security architecture, runtime defenses, and governance controls for Replit’s AI agents and their code execution, tool use, and data access.
What you'll do
- Define security architecture for agent workflows, integrations, and coordination; build guardrails and sandboxing for agent execution.
- Threat-model and red-team agent systems, and establish controls for identity, permissions, data isolation, logging, and risk tracking.
- Document secure design patterns and partner with engineering, GRC, and Sales on enterprise audit and security needs.
What they ask for
- 6+ years in security engineering or architecture, including at least 2+ years dedicated to AI/ML security, LLM application security, or agentic framework security.
- Understanding of agentic architectures and protocols, and hands-on expertise with agentic AI threat vectors.
- Experience applying AI safety standards and risk frameworks, and proficiency in Python, Go, or TypeScript/JavaScript.
- Experience authoring security architecture documentation and contributing to an enterprise Cybersecurity Risk Register.
Nice to have
- Experience designing runtime sandboxing and isolation technologies such as Firecracker, gVisor, Docker, or WebAssembly.
Pay
$230 – $350 • Offers Commission
Tools
Model Context Protocol (MCP), LangChain, AutoGen, CrewAI, Python, Go, TypeScript/JavaScript, WebAssembly
AI focus
Secures autonomous AI agents through runtime guardrails, sandboxing, threat modeling, red teaming, least-privilege controls, data isolation, and auditability.
Model-written summary · the original posting is authoritative