AI Solutions Engineer - Ken Garff Automotive Corporate Offices - SLC, UT
Position Summary
The AI Solutions Engineer will help business teams build their own apps with Claude Code, safely and inside company boundaries, and get them deployed. You pair program with the people closest to the problem, from a parts manager or service technician up to an executive vice president. The business unit owns the app. You make sure it is built securely, deployed correctly, and ready for a clear decision on its future. The work is almost entirely Claude Code and vibe coding. It suits you if you enjoy building quickly, teaching as you go, and knowing when to slow down. Location and interviews This is a hybrid role based in our Salt Lake City office. With the exception of an initial HR phone screen, all interviews will be conducted in person at our Salt Lake City office - no video interviews for subsequent rounds. Relocation assistance is not available; candidates must already be located in the Salt Lake metro area or be willing to relocate at their own expense before starting.
Key responsibilities
- Help with intake: point each requester to the inventory checklist before any building starts, and help keep the app inventory current.
- Pair program with business users of any technical level, using Claude Code to build working apps.
- Build within company boundaries: handle data, credentials and access the way our security and AI governance standards require.
- Keep all code in Git and through code review.
- Deploy apps.
- Support apps during business hours only. Anything that becomes mission critical moves to the enterprise tier.
- Think through architecture twice: for the quick deployment, and for what it would take to productionize the app later with the main team.
- Resume pairing with the business owner when an app breaks.
- Help run the post-launch and yearly reviews, and recommend whether to retire an app, promote it to the enterprise level, or keep it small. The same choice applies if an app’s owner leaves, or you help find a new owner.
- Tell business users plainly when an idea is not worth building, or should not be built a certain way.
- Document each app so someone else can pick it up.
- When work is light, spend a sprint on the main AI team.
Required qualifications The people part comes first, because it is the hard part of this job.
- Excellent communication and teaching skills. You can pair with a service technician in the morning and an executive in the afternoon, and make both feel capable.
- Good judgment and a willingness to push back respectfully when an idea is unsafe or not worth building, while staying someone people want to work with again.
- Able to read, review, and explain the code Claude writes, and debug it when it breaks.
- Strong Claude Code and AI-assisted development skills. I want to see apps you have actually built and shipped.
- Comfort with Git and code review.
- Working knowledge of application security: secrets handling, access control, and protecting sensitive data such as customer information.
- Basic architecture sense: able to tell a quick internal tool from something that must scale, and to explain the difference.
Nice to have None of these are required.
- Experience building MCP servers or LangChain deployments.
- Background in a regulated industry or with compliance requirements.
- Experience with dealership operations or automotive retail.
- Some HTML, CSS, and JavaScript.
- Basic SQL, ideally with Snowflake.
- Familiarity with Domo, including Pro Code Apps.
How I hire There is no ATS on this listing. I’m the one reading your resume, and I will ask you about anything on it. I expect you to use AI. If you give it this posting, tell it: no keyword stuffing, don’t mirror the posting back to me, and only include what you can talk about in detail. I want to see the human behind the resume. When you apply Skip the generic cover letter. Instead, submit your answer to one of these two questions as your cover letter, in your own words and in a few short paragraphs:
- Tell me about a time you helped someone who was not technical build or fix something real. Then tell me about a moment it got hard, such as when they wanted something you thought was a bad idea. What did you do?
- Tell me about one thing you built with AI that broke, and what you did about it.
Pick one. Not both, and not neither. If you have something I can look at, such as a link, a demo, or a screenshot, include it. I read these.