Job Summary
The AI Engineer will develop, configure, and optimize coding agent solutions, with a focus on API development, agent configuration, output quality, and automated verification. This role requires an experienced API developer with hands-on expertise in coding agents and spec-driven development methods, rather than a prompt-focused specialist. The engineer will establish reusable templates, define output standards, validate generated code, and continuously improve agent performance based on failures and reviewer feedback.
Key Responsibilities
• Configure headless coding agents for different archetypes, including steering files, agent configurations, Model Context Protocol (MCP) context, and result contracts.
• Build the first golden template to establish standards for the remaining four templates.
• Develop output verification mechanisms, including path allowlists, contract tests, and review-ready pull request structures.
• Run iterative pilot tuning cycles by analyzing failures and reviewer feedback, then improving templates and steering configurations.
• Collaborate with the client's API developer to develop templates 2 through 5.
• Establish and maintain standards for generated code quality, consistency, and review readiness.
• Evaluate agent behavior and identify failure modes to improve reliability and output quality.
Required Qualifications
• 5+ years of experience building production APIs using REST and at least one of GraphQL or event-driven architecture on AWS.
• Hands-on experience with coding agents such as Kiro, Claude Code, or similar tools, including spec-driven or headless usage beyond IDE autocomplete.
• Strong testing discipline, including experience with contract testing and OpenAPI.
• Ability to assess generated code quickly and determine its quality, correctness, and readiness for review.
• Ability to clearly explain coding agent behavior and failure modes to reviewers and clients.
Preferred Qualifications
• Experience building reusable templates, scaffolding, or code generators adopted by other teams.
• Experience with agent evaluation, including measuring acceptance rates and comparing configurations.
• Experience with BMAD or similar spec-driven development methods.
