Prompt Engineer - Enterprise AI

White Tree Solutions

White Tree Solutions is seeking a Prompt Engineer to design, test, and refine AI interactions for enterprise use cases. This role combines clear instruction design with systematic evaluation to help AI solutions produce useful, grounded, and consistent responses.

You will collaborate with architects, business analysts, governance specialists, and end users to translate business requirements into tested AI behavior.

• Develop system instructions, prompt templates, examples, and response formats for approved use cases.
• Define how AI solutions should use supplied context, retrieve information, handle ambiguity, and elevate unsupported requests.
• Build representative test cases and evaluation criteria covering accuracy, relevance, instruction adherence, and appropriate refusals.
• Evaluate behavior under ambiguous questions, incomplete context, conflicting instructions, and adversarial inputs.
• Diagnose failure patterns and recommend changes to prompts, source content, retrieval, or application design.
• Maintain versioned prompt documentation, test results, and release recommendations.
• Collaborate with security and engineering teams on safeguards that require controls beyond prompt instructions.

• Hands-on experience building and evaluating applications using large language models.
• Strong writing skills and the ability to express complex requirements as clear instructions.
• Experience developing repeatable tests and analyzing AI outputs.
• Understanding of grounding, retrieval-augmented generation, context limitations, and common model failure modes.
• Ability to work with technical teams and explain tradeoffs to business stakeholders.

• Microsoft Foundry, Copilot Studio, or Microsoft Fabric data agents.
• Python, APIs, and automated evaluation workflows.
• Conversational analytics or enterprise knowledge assistants.
• AI solutions subject to formal governance and release controls.

A documented prompt library, evaluation dataset, test results, and recommendations for improving AI behavior and release readiness.

Candidates must be U.S. citizens. This role supports a government engagement that prohibits offshoring. All work, including access to client systems and data, must be performed within the United States throughout the engagement. Offshore delivery and subcontracting are not permitted.