Founding AI Engineer | Seed-Stage FinTech | Up to £160k

Few&Far; Greater London; Vor Ort; GBP 96.000 - 160.000; Vollzeit

Benefits dieser Stelle

Equity

Zusammenfassung

Few&Far in London is seeking a founding AI engineer to own how AI actually does the audit work, from first prototype to a trusted client-ready product. You will build agentic workflows over client accounting data, integrate with ERPs, banks and invoicing tools, and shape the platform end to end.

You will work closely with auditors, ship end-to-end, and ensure every step is auditable with clear source attribution. This is an onsite role in London.

Qualifikationen

• Real LLM depth, e.g. agent harnesses or LLM product engineering.
• Shipped projects or side projects count.
• Daily AI work focus, not just an add-on.
• Strong fundamentals over specific tools.

Aufgaben

• Build agentic workflows over real client accounting data, from the harness and evals to what auditors see.
• Take problems end to end, decide how to solve them and ship them.
• Build the control and evidence layer, so every step is traceable to its source and a named auditor.
• Sit with the auditors using your work every week and iterate fast.
• Get stuck into messy, real integrations with ERPs, banks and invoicing tools.

Kenntnisse

LLM depth

Shipping experience

AI product engineering

Fundamentals

Jobbeschreibung

Founding AI Engineer| AI Audit Start-up | London ([onsite) | Up to £160k + equity

Audit is one of the last big professional services that AI hasn't properly cracked yet. This team is going after it.

I'm working with a London start-up rebuilding audit for small and mid-sized companies, a market the big firms are quietly walking away from. Rather than selling software to auditors, they do the whole thing themselves: they hold the licence, build the platform and employ the auditors who sign off. Their product plugs into a company's ERP, banks and invoices, and picks up each year where the last one finished.

They're hiring an engineer to own how AI actually does the audit work, from first prototype to something clients rely on.

What you'll do

• Build agentic workflows over real client accounting data, from the harness and evals to what auditors see
• Take problems end to end, decide how to solve them and ship them
• Build the control and evidence layer, so every step is traceable to its source and a named auditor
• Sit with the auditors using your work every week and iterate fast
• Get stuck into messy, real integrations with ERPs, banks and invoicing tools

What they're looking for

• Real LLM depth, whether that's agent harnesses, LLM product engineering or similar
• Stuff you've actually shipped (side projects count)
• Someone who builds with AI every day, not as an add-on
• Strong fundamentals over specific tools

Why it's worth a look

• A big, unglamorous market that's ripe for change, and they own the whole stack to change it
• The hard problem is trust and evidence, not just prompting a model
• Your work goes straight in front of real users every week