Principal AI Engineer

Principal AI Engineer
Location: Columbus, Ohio
Duration: 6-12 Months
Job Role:
You'll be the founding engineer, building and operating the AI platform and the first production applications on it — end to end, from infrastructure and orchestration through evaluation and monitoring. The engineering standards you set — repository structure, testing, deployment patterns — become the foundation every future hire builds on. This is a principal-scope role with a genuinely hands-on charter: you ship code every week and own it in production.
What you'll do
Build and operate the governed enterprise MCP gateway and Azure AI environments
Implement SAP identity passthrough in partnership with the IT/SAP organization
Own CI/CD for AI assets, plus the evaluation and observability harness with cost and usage telemetry instrumented from day one
Ship the first production AI use case, then the prioritized portfolio: retrieval, grounding, and agentic workflows, released behind evaluation gates
Co-build alongside an enterprise integration partner during the initial platform build, internalizing patterns so the capability stays in-house
Define the function's engineering standards, tooling, and practices, and serve as technical lead as the team grows
Your first 90 days
Have the gateway carrying production traffic for the first use case in a Hexion-owned Azure environment
See evaluation scores and telemetry flowing from day-one instrumentation
Publish the engineering standards the team will build on
Required qualifications
10+ years of software engineering; strong Python and/or TypeScript
Deep Azure expertise: infrastructure-as-code (Bicep or Terraform), networking, identity, container platforms, Azure AI services
Production LLM or agentic systems you personally built and ran end to end — data flows, orchestration, deployment, evaluation, incident response — be prepared to walk through the most recent one
Evaluation engineering: you design and build eval suites — datasets, metrics, regression tests wired into CI — and hold them as release gates, not demos
Ownership of CI/CD pipelines (GitHub Actions, Azure DevOps, or equivalent)
Identity-aware API and integration design
Sound judgment partnering with vendors while keeping core capability in-house
AI-augmented development as your default — we work with Claude and Microsoft Copilot, and expect them in your daily build loop
Preferred qualifications
Model Context Protocol (MCP) implementation experience
SAP or other enterprise ERP integration exposure
Experience with enterprise integration/automation (iPaaS) platforms
Prior technical-lead or mentoring experience
Regulated or industrial-sector environments

Work arrangement
No