Security Software Engineer, Principal (AI)

TL;DR

Security Software Engineer, Principal (AI Security): Designing and building a secure runtime substrate to detect and contain threats from autonomous AI agents with an accent on sandboxed execution and hard isolation. Focus on implementing microVM-based isolation, real-time behavioral detection, and policy-driven response orchestration for enterprise AI systems.

Location: Hybrid in Bellevue, Washington

Company

hirify.global is a global leader in CRM and enterprise cloud software, currently integrating autonomous AI agents across its platform.

What you will do

• Design and own the secure execution substrate for AI agents using microVM-based isolation technologies like Firecracker and Kata Containers.
• Build a runtime detection layer to score agent behavior against learned baselines and detect anomalies in real time.
• Implement a policy enforcement and response orchestration layer to execute reversible containment actions (throttle, suspend, revoke).
• Develop schema-governed telemetry pipelines using OpenTelemetry and Prometheus to capture end-to-end agent traces at scale.
• Create evaluation and replay infrastructure to test detection models and containment policies against historical traces.
• Maintain platform infrastructure as code using Terraform and CI/CD pipelines via GitHub Actions and ArgoCD.

Requirements

• 9+ years of experience as a Platform, Security Infrastructure, or Software Engineer in high-trust production environments.
• Hands-on expertise with microVM and container isolation (Firecracker, Kata Containers, gVisor).
• Experience building scaled sandboxed code-execution environments for AI agents.
• Strong proficiency in Python and JVM-based languages.
• Deep expertise in cloud platforms (AWS, GCP, or Azure) and orchestration (Kubernetes).
• Background in streaming systems (Kafka, Flink) and production ML serving.

Nice to have

• Experience with AgentOps, agentic loops, and tool orchestration.
• Familiarity with agent evaluation frameworks like LangSmith or OpenAI Evals.
• Background in offensive security, red-teaming AI systems, or sandbox escape research.
• Experience with vector/graph databases and RAG pipelines.

Culture & Benefits

• Office Tech-Flexible hybrid work environment.
• AI-first engineering culture with heavy use of tools like Claude Code, GitHub Copilot, and Cursor.
• Commitment to enterprise-grade security and compliance (SOC2, FedRAMP, ISO).
• High-impact role working on the cutting edge of autonomous AI agent safety.