Note: The job is a remote job and is open to candidates in USA. Feuji is seeking a Staff Engineer I Agentic AI Platform Deployment Engineer to take autonomous AI agents from validated design into safe, observable production operation. The role owns deployment pipelines, security boundaries, adversarial testing, human-in-the-loop controls, observability, incident response, and collaboration on rollout strategy.
Responsibilities
- Own the deployment pipeline: staging, canary rollout, and production cutover
- Define least-privileged permission and tool access boundaries for autonomous agents
- Run pre-launch adversarial testing (prompt injection, tool misuse, unsafe plans) before production exposure
- Implement human-in-the-loop approval gates and escalation paths for high-risk agent actions
- Build observability, audit logging, and kill switch/rollback mechanisms for agent behavior
- Lead incident response for agent failures and track task success, cost, and latency in production
- Partner with architects, AI engineering, and security on rollout strategy; mentor engineers on safe deployment
Skills
- Bachelor's degree in computer science or equivalent
- 8 to 10+ years enterprise software engineering experience, with ownership of production deployment/operations
- Hands-on experience deploying or operating agentic AI systems (Claude, Copilot, Edifecs, or Google's agentic platform)
- Strong grasp of AI safety and guardrail patterns: sandboxing, permission scoping, rollback design
- Practical CI/CD, DevSecOps, and cloud deployment experience (OpenShift, Kubernetes, AWS)
- Direct experience with Edifecs Agent Platform or Google's agentic platform in production
- Experience running red team or adversarial evaluations against LLM agents
- Experience defining SLOs and alerting for agent reliability
Company Overview
Company H1B Sponsorship