AI Software Engineer

About the role

• Job Posting Title
• AI Software Engineer
• Agency
• 090 OFFICE OF MANAGEMENT AND ENTERPRISE SERV
• Supervisory Organization
• IS-Data
• Job Posting End Date

Refer to the date listed at the top of this posting, if available. Continuous if date is blank.

Note: Applications will be accepted until 11:59 PM on the day prior to the posting end date above.

• Estimated Appointment End Date (Continuous if Blank)
• Full/Part-Time
• Full time
• Job Type
• Regular

About the role

As an AI Software Engineer with OMES you will enjoy

• Generous leave including 15 days of vacation, 15 days of sick leave and 11 paid holidays annually.
• A comprehensive Benefit Package with a generous benefit allowance to offset the cost of insurance premiums for employees and their eligible dependents.

Job Details

• Full-time 40-hour work weeks
• Support the Information Services Division
• Salary is based on education and experience
• This is an on-site position located in Oklahoma City, OK

Position Summary

The AI Software Engineer owns the agency's custom AI development on Microsoft Foundry (formerly Azure AI Foundry). This is a hands-on engineering role with a clear mandate: turn AI delivery from a series of one-off projects into a repeatable, industrialized pattern that lets the organization deliver more AI solutions, faster, at a lower cost per solution.

The role inherits AI products already running in production. The engineer will take full ownership of them - stabilizing, hardening and supporting the code - while extracting what works into a shared core capability that every future AI solution is built on. Success means each new use case is a configuration of proven components, not a new build from scratch.

Key Responsibilities

Reusable AI Platform & Delivery Pattern (Economy of Scale)

• Design and maintain a reference architecture for AI solutions on Microsoft Foundry, including Foundry Agent Service, Foundry Models, Azure OpenAI and Azure AI Search / Foundry IQ grounding.
• Build reusable starter templates, shared SDK/libraries and infrastructure-as-code modules so new use cases can be stood up in days rather than months.
• Standardize common building blocks: prompt and agent configuration, retrieval-augmented generation (RAG) pipelines, tool/API connectors, model routing, conversation state and error handling.
• Define a repeatable intake-to-production lifecycle (discovery, prototype, evaluation, security review, deployment, support handoff) and continually refine it based on lessons learned.
• Identify duplicated logic across products and refactor it into the shared core; retire one-off code wherever possible.
• Track and optimize unit economics - token consumption, model selection, compute and hosting cost - and design for cost efficiency at scale.

Inherited Product Ownership & Code Hardening

• Assume technical ownership of existing production AI products; perform an early assessment of architecture, code quality, security, dependencies and technical debt.
• Develop and execute a prioritized hardening plan: automated unit, integration and regression tests; static and dependency scanning; secrets management; resiliency and retry patterns; rate-limit and quota handling.
• Migrate inherited products onto the shared core capability and standard pipelines as the pattern matures.
• Implement automated AI evaluation (groundedness, relevance, accuracy, safety) as release gates, and maintain regression test sets for each product.
• Deliver enhancements and new features against a prioritized backlog in partnership with product owners and business stakeholders.

Production Operations & Support

• Provide ongoing support for AI products, including participation in an on-call rotation and incident response for production issues.
• Instrument products with telemetry, tracing and dashboards (Azure Monitor, Application Insights, Foundry observability) covering latency, error rates, cost and model quality.
• Define service level objectives with stakeholders and monitor model and prompt drift, degradation and anomalous usage.
• Perform root cause analysis, write post-incident reviews and drive corrective actions back into the shared platform.
• Manage model version lifecycles, deprecations and upgrades with minimal disruption to users.

Security, Governance & Responsible AI

• Apply secure-by-design practices: managed identities, least-privilege RBAC, private networking, key and secrets management, and data protection appropriate for government data.
• Implement content safety, prompt-injection defenses, input/output filtering and human-in-the-loop controls where required.
• Support the agency's AI governance framework, documentation and audit requirements, including model cards, data lineage, risk assessments and records retention.
• Ensure solutions meet applicable state policy, privacy, accessibility (WCAG / Section 508) and security standards.

Collaboration, Enablement & Documentation

• Work within an agile delivery model alongside DevSecOps, SRE, Observability, infrastructure, security and data teams.
• Produce clear architecture documentation, runbooks and developer guides so the pattern can be adopted by other engineers and vendors.
• Coach developers and contractors on the standard pattern; perform code reviews and enforce engineering standards.
• Translate business problems into technical approaches and communicate trade-offs to non-technical stakeholders and leadership.

Physical Demands and Work Environment

• While performing the duties of this job, this position works in an office setting with a computer for a large percentage of the workday.
• The noise level in the work environment is usually moderate.
• Occasional travel may be required.

This position participates in an on-call rotation for production AI services, including occasional after-hours support.

Minimum Qualifications

Requirements include a Bachelor's degree in Computer Science, Software Engineering or a related field and 6+ years of professional software development experience building and supporting production applications, plus 2+ years developing generative AI or LLM-based solutions in production, including RAG and/or agent-based architectures, or equivalent combination of education and experience.

Preference will be given to candidates who possess

• Microsoft Certified: Azure AI App and Agent Developer Associate (AI-103) or predecessor Azure AI Engineer Associate (AI-102); Azure Developer or DevOps Engineer certifications.
• Experience with Foundry Agent Service, Semantic Kernel / Microsoft Agent Framework, multi-agent orchestration or Model Context Protocol (MCP) tooling.
• Experience with AI evaluation frameworks, LLMOps practices and prompt/version management.
• Background in public sector, regulated industries or environments with strict security and compliance requirements.
• Experience with FinOps or IT cost management for cloud and AI workloads.
• Familiarity with Power Platform, Copilot Studio or Microsoft 365 integration.