Senior AI Solutions Engineer

Reports to: EVP, Technology & Innovation Location: On-site in Columbia or Kansas City, Missouri preferred. Remote considered for the right candidate.

What you'll actually do

This role spans four kinds of work. You'll move between them constantly.

1. Get more out of the tools we already have. We run Microsoft Copilot, Claude, and ChatGPT across the enterprise. Most people use a fraction of what these can do. You'll build agents, projects, and workflows on top of them, write the prompts and guardrails that make them reliable, and help ~3,000 professional staff actually adopt them.

2. Connect AI to our data. Our analytical source of truth is an enterprise data warehouse, fed from our ERP and other source systems. Our documents live in cloud file-sharing and collaboration platforms. You'll wire AI tools into these sources - via APIs, connectors, and retrieval pipelines - and expose them as agents and projects people can use without a data engineering degree.

3. Evaluate and drive commercial (COTS) solutions. Not everything should be built. You'll assess vendor AI products, run honest build-vs-buy analyses, and when something makes sense, drive the implementation across the relevant subsidiaries. This requires the credibility to sit in an operations meeting and be believed.

4. Build custom automation and applications. When there's no good tool to buy, you build it. Our stack is Google Cloud Platform (Cloud Run, Cloud SQL/PostgreSQL, Pub/Sub, Cloud Scheduler, Secret Manager), BigQuery and dbt for data, and a CI/CD-driven monorepo. AI shows up two ways here: sometimes it's in the product (extraction, classification, agents), sometimes it's just how you build faster (we use Claude Code heavily), and often both. You're expected to be fluent in agentic development workflows, not just aware of them.

What we're looking for

• You can actually build. You write production code (Python and SQL at minimum), deploy it, and maintain it. You're comfortable in the cloud - GCP preferred, but we'll take strong AWS/Azure if you can switch.
• You're fluent with modern AI tooling: LLM APIs, prompt design, retrieval/RAG patterns, agent frameworks, and AI-assisted development (Claude Code, Copilot, or equivalent). You understand where these tools are reliable and, more importantly, where they aren't.
• You can work with data: SQL is second nature, you understand warehouses and ELT, and you can model and query messy real-world business data.
• You can translate. You can talk to a project manager about job costing, then go write the code, then explain the result to a CFO. The translation is half the job.
• You have good judgment about build vs. buy, and you're not religious about either.
• You operate with minimal direction. You can take a vague business problem, scope it, and come back with something that works.

Nice to have

• Construction, engineering, or other operations-heavy industry experience.
• Experience integrating with enterprise systems such as ERP (Viewpoint Vista or similar) and HCM platforms (Workday).
• dbt, BigQuery, and modern data stack experience.
• Experience standing up internal-facing applications with enterprise auth (Entra ID / OIDC).
• A track record of getting non-technical people to adopt new tools - adoption is harder than building.

Who thrives here

The person who does well in this role is pragmatic over flashy, ships over theorizes, and would rather deliver a credible $100K improvement than promise an unsupportable $1M one. You're comfortable being early, you don't need a playbook handed to you, and you can hold your own with operators who are skeptical of anything with "AI" in the name.