AI / ML Engineer
Take models from evaluation into production, with the confidence scoring, review paths and monitoring that let a client trust the output.
About the role
Our AI work runs inside client environments, often under residency and audit constraints. The interesting part of the job is not the model — it is making the output something a reviewer will sign their name against.
What you will do
- Build the evaluation set before building the model
- Develop, fine-tune and evaluate models against agreed measures
- Design confidence scoring and the human review path
- Deploy inside the client’s own tenancy with proper MLOps
- Instrument for drift and failure, and act on what it shows
- Write the model card: what it does, where it fails, who reviewed it
What we are looking for
- Python and the modern ML / LLM toolchain in real use
- Can design an evaluation that would catch your own mistakes
- Understands retrieval, prompting and fine-tuning trade-offs
- Takes governance and traceability seriously
Useful, not essential
- MLOps and model deployment at scale
- Document extraction or clinical/legal NLP
- Published work or open-source contributions
How hiring works
A first conversation, then a short practical exercise related to the job, then a conversation with the person you would report to. We reply to every application, and we do not ask for unpaid work.
Apply for this role
Fields marked with an asterisk are required. Attach your CV as PDF, DOC or DOCX, up to 5 MB.