SOX AI Specialist, Fieldguide

• At least four years of hands-on control testing experience in SOX compliance, internal audit, or internal controls, with direct responsibility for control design and testing
• Broad experience across a range of control types, industries, and firms, with clear expertise in business process control testing
• The ability to consistently explain why a test step, exception, or conclusion passes or fails and to turn that reasoning into clear evaluation criteria
• Strong written reasoning: you can explain workflows and decisions step by step and turn professional judgment into clear rules, edge cases, and review criteria
• A high degree of comfort and hands-on experience using AI tools, writing prompts, and working with agentic harnesses
• Comfort operating in ambiguity, testing ideas quickly, and changing your approach when the evidence does not support it
• Availability for the expected time commitment during the initial engagement

Desirable

• We are seeking an experienced SOX and internal controls practitioner for a paid, short-term contract engagement. This is not a full-time role (at this time)
• Over an initial engagement of 8–12 weeks, with the potential to extend, you will work directly with a small, cross-functional group of subject-matter experts, AI engineers, and product partners. You’ll help translate real-world SOX testing judgment into reliable, AI-assisted workflows
• This is not a traditional client-delivery audit or attestation role. Your focus will be on prompting and refining AI agents to execute control testing and partnering closely with AI engineers to identify new engineering capabilities to enable better control testing
• Initial term: 8–12 weeks, with potential to extend
• Partner closely with engineering, product, and design teams to translate real-world SOX testing processes into scalable product capabilities
• Co-develop AI agents with AI engineers to execute detailed SOX testing by translating SOX testing methodology into specific guidance, decision rules, and examples
• Work hands-on with AI agents: write and refine prompts, run controls and evidence through testing workflows, evaluate outputs, and document failure modes
• Analyze risk and control matrices and test plans to identify opportunities for improved agentic testing
• Develop stress-test scenarios, review and assess AI agent performance, and create detailed error analyses and evaluation criteria