Principal AI Application Engineer

Code for America believes in transforming government services through technology, and they are seeking a Principal AI Application Engineer to bridge the gap between policy mandates and technical execution. This role involves architecting modular systems for AI deployment in government agencies, ensuring responsible and effective service delivery.

Responsibilities:

• Build and ship high-impact experiments (e.g., digital lockers or AI-augmented procurement tools) to expand the "impact possibilities frontier."
• Translate vague policy objectives into robust, working systems
• Identify when to use LLMs, agentic orchestration, or RAG patterns versus simple rules engines or traditional ML models
• Collaborate with policy and domain experts to co-design civic-sector benchmarks, translating complex regulatory requirements into automated evaluation pipelines that rigorously measure model performance
• Identify and define new opportunity spaces by translating emerging policy, technology, and user needs into actionable technical bets
• Architect foundational tools, including declarative task specifications and agentic data layers
• Design infrastructure that respects public sector constraints, focusing on portability and explainability
• Build data layers that interoperate with legacy systems (COBOL, SQL, etc.) to deliver modern value without multi-year migrations
• Establish Code for America as a leader in responsible AI through external thought leadership, including publications, talks, and open-source contributions
• Share demos and earned insights internally to help the organization iterate toward better standards and internal use cases of responsible AI
• Document architectural decisions, successes, and failures to create a blueprint for responsible AI in government
• Drive alignment across engineering, product, policy, and program teams to ensure solutions are technically sound, policy-compliant, and operationally viable
• Partner with and mentor fellow engineers through hands-on code reviews and technical guidance, ensuring the team stays grounded in best practices for responsible AI
• Maintain system health through rigorous, hands-on code reviews and the development of shared utilities
• Ensure craftsmanship and system explainability for the vulnerable populations we serve

Requirements:

• 7+ years of experience in high-ownership environments (former technical founders encouraged); ability to take a vague objective to a finished system
• Hands-on experience building with LLMs, agentic orchestration, and RAG patterns, with the pragmatism to know when not to use them
• Ability to think in 'primitives' and 'capabilities,' preferring modular, reusable frameworks over bespoke scripts
• Passion for bias detection, harm mitigations, and building systems that are explainable to the people they serve
• Mastery of Git, Linux, CI/CD, Infrastructure as Code (Terraform), and container-based workflows
• A critical eye toward the limitations of AI, especially 'black box' logic in high-stakes public services
• Demonstrated ability to influence technical direction and drive alignment across teams without formal authority
• Ability to distill complex architecture into compelling prose for policy-makers or the public
• Prior work in Civic Tech, FinTech, or HealthTech where auditability is a core requirement