Open role
LLM Engineer
Build LLM-assisted systems for regulated finance. Every output must be testable, traceable and fit for a person to sign off.
About the role
Build what can be verified.
We are looking for an experienced software engineer who has put LLM systems into production. The work combines conventional engineering with probabilistic components. You must know where each belongs, how each fails and what evidence a reviewer needs.
You’ll work independently within controls aligned to FCA expectations and ISO 42001. We want evidence of production work: evaluations, failure handling, human review and the audit trail behind a release.
We grow slowly. We need engineers whose judgement remains sound when the model’s output is plausible and wrong. Read our AI Policy for more on how we work.
The work
What you will own.
- Design and implement LLM-assisted financial applications, with human review and sign-off at every release stage
- Build DSPy and prompt workflows for business rules that can be tested and audited
- Develop memory systems and AI agents using mem0 and other appropriate frameworks
- Create serverless architectures and automation workflows using N8N
- Implement MCP integrations with explicit permissions, failure states and audit trails
- Apply test-driven development to AI-assisted applications, with particular attention to edge cases and regulatory requirements
- Maintain audit trails documenting where AI has been used in the development process
- Verify every AI-generated output before it enters a deliverable, including code, documentation and analysis
What you bring
The groundwork.
- 5+ years of professional software engineering experience with production systems
- Production experience with DSPy or an equivalent framework for structured LLM applications
- Proficiency with MCP (Model Context Protocol) and AI system integration
- Strong React, TypeScript and serverless architecture foundations
- Experience with N8N, mem0 and test-driven development for AI systems
- Deep understanding of output verification, human review and the failure modes of probabilistic systems
- Awareness of data protection requirements (UK GDPR) and regulatory standards (FCA, ISO 42001) as they apply to AI in financial services
Show your work
Show us where it ran.
- Demonstrable evidence of advanced LLM usage in previous projects
- A portfolio showing AI-agent development and prompt-engineering work
- Examples of DSPy implementations and serverless AI applications
- Experience working within AI governance frameworks, or a clear understanding of responsible AI practices in regulated industries
In return
What we provide.
- Salary from £45,000 a year, based on relevant experience
- Flexible hours and remote work within the UK
- Professional development budget
- Health insurance
- 25 days holiday plus bank holidays
How to apply
Send your work.
Send a CV and one example of an LLM system you have put into production. Explain what could fail, how you measured it and where a person remained accountable. Write to .
Note
- This is a remote position for applicants based in the UK.
- We hire individual engineers, not outsourced teams.
- No recruiters or agencies, thank you.
Apply
Build systems a person can sign off.
Bring us production evidence, clear judgement and work you can answer for.