Prompt Engineer

We seek an AI Prompt Engineer with strong analytical skills, excellent reading comprehension, and exceptional attention to detail. You will design, test, and refine prompts for large language models, audit AI-generated outputs for accuracy, and work closely with client teams to ensure AI solutions meet production quality standards.

• Design and refine structured prompts for AI models to extract data and generate narrative outputs from complex source documents.
• Identify and diagnose AI failure modes such as fabricated data, incorrect source attribution, and formatting errors, and develop targeted prompt fixes.
• Iterate prompts through structured test-and-audit cycles to meet defined accuracy thresholds.

• Audit AI-generated outputs against source documents, verifying data points at the individual figure level.
• Build validation workbooks to track accuracy metrics across test runs.
• Categorize and document errors by type and severity to drive systematic improvement.
• Work with client subject matter experts to understand workflows, quality expectations, and output conventions.
• Translate domain-specific drafting conventions into structured AI instructions.
• Present validation results and accuracy metrics to internal and client stakeholders.

• Education: Bachelor’s degree in Finance, Accounting, Business, Economics, or a related field.
• Experience: 0 to 2 years including internships, academic projects, or full-time roles.
• Strong analytical and quantitative reasoning with comfort working with numbers, percentages, financial figures, and tabular data.
• Exceptional attention to detail with a natural tendency to validate numbers and verify source accuracy.
• Excellent reading comprehension with the ability to interpret dense, data-heavy documents such as financial reports and regulatory filings.
• Clear, precise, and structured writing skills to create unambiguous AI prompts.
• Proficiency in Excel or Google Sheets for data validation and tracking.
• Patience and discipline for methodical, repetitive quality assurance work.
• Strong logical reasoning to trace AI errors to root causes and implement systematic fixes.