Founding Engineer - Autonomous AI Audit Platform (London)

Dex; Greater London; On-site; GBP 90,000 - 140,000; Full time

Job summary

Dex is partnering with a leading European company to replace traditional audit workflows with autonomous, self-improving agent systems. The founding engineer will own the agent layer that plans, executes, and checks audit work against international standards, collaborating with licensed auditors.

You will turn ambiguous, one-sentence domain problems into dependable production-ready architectures and drive them from concept to live systems alongside the core team in London.

Qualifications

• Track record of owning and shipping production software end to end.
• Hands-on experience building, evaluating, and operating LLM systems in production.
• Ability to break down complex, ambiguous real-world workflows directly alongside non-technical domain experts.

Responsibilities

• Architect and ship the agent systems responsible for planning, executing, and evaluating statutory audit workflows against regulatory standards.
• Build the evaluation machinery and feedback loops that allow agentic systems to reliably self-check their evidence trails.
• Interface the agent layer with a deterministic transaction engine to verify every single ledger entry rather than relying on statistical sampling.
• Work directly with state-authourised auditors to translate complex legal rules, materiality judgments, and evidentiary standards into autonomous code.
• Take ambiguous technical problems from initial concept to live production systems without waiting for tickets or detailed roadmaps.

Skills

End-to-end software delivery

LLM systems in production

Ambiguous real-world workflows

London-based

Job description

Dex is partnering with a leading European company to replace traditional audit workflows with autonomous, self-improving agent systems. The founding engineer will own the agent layer that plans, executes, and checks audit work against international standards, collaborating with licensed auditors.

You will turn ambiguous, one-sentence domain problems into dependable production-ready architectures and drive them from concept to live systems alongside the core team in London.