We are building multiple AI-enabled tools to support the audit lifecycle, from document review and control analysis to risk insight generation and follow-up. This internship gives an engineer hands-on experience turning real audit needs into working, well-documented AI products. The intern will contribute to audit assurance tools while learning how to design, build, evaluate, and improve trustworthy applications for auditors.
This project develops LLM-powered tools that help audit teams in various aspects of assurance work:
- Analyse reports, policies, evidence, and control requirements.
- Identify recurring themes, risks, and control weaknesses, and present traceable insights for review and decision-making.
- Automate workflows and processes across audit planning, fieldwork, reporting, and follow-up.
AI will serve as a first-line guide to explain audit requirements, identify relevant evidence, and help users prepare for review. Outputs will be grounded in approved policies and control sources, cite sources where available, and remain subject to professional judgement.
The intern will work across the product lifecycle, from understanding an audit problem and shaping a solution through prototyping, testing, deployment, demonstration, and iteration. The solution must preserve security, explainability, and human oversight.
What You Will Work On
- Translate audit assurance workflows into clear user needs, technical requirements, assumptions, and measurable outcomes.
- Build secure LLM, RAG, and agent workflows that process audit material, generate traceable insights, and automate control review and follow-up with human approval.
- Develop usable full-stack features, including web interfaces, backend services, APIs, databases, dashboards, and export or reporting functions.
- Test AI outputs against representative reports and edge cases; assess relevance, factual grounding, consistency, and usability; and apply security and human-review safeguards.
- Use Git, issue tracking, code review, testing, and deployment practices to deliver maintainable solutions and document key decisions, limitations, and improvements.
