US CPA – AI Tax Review Specialist

Overview

This role involves supporting the development of an advanced system that integrates large language models with the Lean 4 theorem prover to translate tax statutes into precise, machine-verifiable logic. The system also produces plain-English explanations to ensure the original statutory meaning is preserved. Due to the complexity and sensitivity of tax law, even minor logical errors can lead to incorrect outcomes.

The position requires acting as a critical quality gate by reviewing AI-generated outputs for accuracy, completeness, and logical correctness. It is well-suited for experienced tax professionals who enjoy detailed analysis and structured problem-solving.

Responsibilities

• Review AI-generated translations of tax statutes into structured, formal logic to ensure accuracy and completeness.
• Evaluate system-generated test cases to confirm they correctly reflect statutory rules and real-world applications.
• Identify issues such as missing exceptions, incorrect thresholds, logical gaps, or misinterpretations of tax provisions.
• Develop additional test scenarios to capture edge cases and ensure robustness across a wide range of conditions.

Minimum Qualifications

• Certified Public Accountant (CPA – US).
• Hands-on experience with US tax practice.
• Strong knowledge of the Internal Revenue Code (IRC) and Treasury Regulations.
• Ability to interpret complex statutory language and translate it into structured logic.
• Strong analytical thinking and attention to detail.
• Ability to work independently and meet deadlines.

Preferred Qualifications

• Experience with AI/large language model evaluation, annotation, or structured review workflows.
• Exposure to rule-based systems, formal logic, or technical validation processes.

Compensation & Engagement

• Role Type: Freelance / Contract.
• Location: Fully Remote.
• Compensation: $60–$80 USD per hour, based on experience and performance.
• Schedule: Flexible, asynchronous.

Location

This is a fully remote position.