Job Description & Details
This is a hybrid gig in Atlanta bridging traditional relational database management with emerging LLM workflows. You will spend your days keeping MS SQL Server humming while simultaneously figuring out how to make AI models spit out cleaner, more accurate SQL and textual responses.
What You'll Actually Be Doing
Expect to split your time between classic database tuning and modern AI enablement. Half of your day is going to be spent deep in execution plans, writing complex stored procedures, and dealing with indexing strategies to ensure high availability. The other half involves iterating on prompt designs, collaborating with data scientists to hook those prompts into existing systems, and validating that the AI outputs are actually reliable and production-ready rather than hallucinating garbage.
The Core Tech Stack
Your bread and butter here is MS SQL Server—you need to know performance tuning, query optimization, and health monitoring inside and out because slow queries will tank the AI integrations built on top of them. On the AI side, solid grounding in prompt engineering techniques is mandatory. You need to understand context windows, zero-shot vs few-shot prompting, and how to systematically evaluate prompt accuracy so the integration with the database layer actually works seamlessly.
Interview Expectations
Be ready for the hiring manager to throw a complex, poorly performing query at you and ask you to optimize it on the spot while explaining your thought process on indexing and execution plans. They will also likely test your prompt engineering mettle by giving you a vague business requirement and asking you to design a prompt strategy that extracts structured JSON data reliably from an LLM, looking for how you handle edge cases and token constraints.
Application Advice
Skip the generic template resume and make sure your CV clearly highlights both sides of this hybrid role: heavy relational database administration alongside concrete examples of prompt tuning and AI model integration. Weave in keywords like query tuning, stored procedures, zero-shot prompting, and database performance monitoring so your profile clears the ATS and lands directly in the hands of the lead engineer.