Security Platform Automation Engineer

Responsibilities

• Lead the software architecture and agentic engineering of an AI-assisted Security Automation Platform.
• Design and operate AI agentic review systems, workflows, skills, tool layers, and evaluation harnesses.
• Build production automation for security reviews across thousands of backend services.
• Integrate first-party and third-party APIs, tools, services, and security data systems into reliable end-to-end workflows.
• Partner with security engineers and analysts to convert review standards into running software and validate output quality.
• Develop bounded-autonomy workflows, human-in-the-loop feedback loops, and high-throughput tooling that reduces reviewer toil.

Requirements

• At least 1 year of experience building LLM- or agent-based systems that ran in production with a measured quality bar.
• 8+ years of experience building and operating production automation, data pipelines, or backend services.
• Highly proficient in at least one of Python, Go, or Java.
• Experience integrating APIs, tooling, and services into reliable workflows and AI agent workflows.
• Working knowledge of application and cloud security concepts, including threat modeling, vulnerabilities, and secure configuration.
• BS or MS in Computer Science, Engineering, or a related field, or equivalent practical experience in Software or Security Engineering.
• Experience with agent engineering concepts such as retrieval, grounding, citation, and reasoning.
• Experience designing agentic workflows with bounded autonomy and evaluating output quality using precision, recall, and human-in-the-loop feedback.
• Familiarity with AWS cloud resources such as S3, EC2, Lambda, and Step Functions.
• Experience with SAST/SCA scanners, SBOM tooling, or vulnerability intelligence platforms.
• Background in security review, threat modeling, or supply-chain security.
• Experience delivering tooling used by non-engineer reviewers or contractors at scale.
• Awareness of AI-specific threats in systems that ingest untrusted input.