Finance & Accounting Expert

Build realistic finance and accounting evaluations

Create benchmark tasks that test whether AI systems can complete real finance and accounting work. Translate practical workflows into demanding assignments with clear reference materials and objective grading criteria. Prior AI experience is not required; the emphasis is on direct professional knowledge and sound judgment.

Responsibilities

• Design tasks involving FP&A analysis, financial reporting, audit procedures, and reconciliations, with multiple steps and enough context for a reviewer to evaluate the result.
• Write clear task instructions and prepare or source the supporting spreadsheets, general ledger extracts, financial statements, and policy documents.
• Define grading criteria that let two qualified reviewers reach the same conclusion about the same output.
• Test the ability to interpret the supplied materials and make judgments, rather than simply retrieve information.
• Identify ambiguous tasks or outputs that cannot be graded reliably, and flag them instead of submitting them as finished work.
• Deliver approved tasks on schedule and meet the required quality standard.

Relevant experience

• At least five years of direct responsibility in FP&A, financial reporting, audit, reconciliations, or accounting. Experience must include doing and owning the work, rather than only supervising it or working alongside it.
• Experience developing or applying structured standards such as audit programs, review checklists, close checklists, grading rubrics, or training assessments.
• Strong Excel skills and the ability to reason across multiple documents. CPA, CA, ACCA, CMA, or equivalent certification is helpful but optional.
• Strong written English; the opportunity lists B2 advanced or C1 minimum English proficiency.
• AI or large language model evaluation experience is welcome but is not a prerequisite.

Contract and selection

The contract is fully remote, with no geographic restrictions. Pay is up to $90 per hour equivalent, and a task is expected to take approximately ten hours. This is project work paid per approved task.

Selection includes screening questions and an AI interview specific to this role, followed by onboarding. The interview presents your domain experience for the client's review. A profile review is also part of the selection process.