Design-Time Threat Modeling: collaborate with product/platform engineers to map new agent types, multi-agent workflows, and AI toolkits; shape trust boundaries and threat models at design time.
Agentic Security Primitives: define primitives for enterprise-scale safety, covering delegation models, multi-agent trust chains, data minimization in RAG workflows, and identity boundaries across customer-built agents.
Architectural Hardening: own security outcomes for complex areas, including hardening Kubernetes or AWS Bedrock controls and designing trust boundaries for MCP integrations; address zero-day risks.
Golden Path Engineering: design policy-as-code and CI/CD controls for agent-assisted workflows to create secure defaults that accelerate engineering without bypassing security.
Structural Improvements: lead detection, response, and architectural mutations after control failures to prevent recurrence.
技術スタック
必須スキル
5+ years of security engineering/architecture at a senior level
Deep understanding of cloud-native architecture (identity, networking, data protection), AWS or equivalent
Production-grade Python, Terraform, or TypeScript for prototype controls and policy checks
Automation-first mindset; prefer system/agent automation over repetitive manual tasks
Ability to earn trust through rigorous logic and influence
歓迎スキル（該当する場合）
Kubernetes hardening and cloud security controls
MCP integrations and multi-agent security design
Policy-as-code tooling and secure-by-default CI/CD patterns
AI/ML security experience or agent-based security background