AI Product Engineer

Prototype useful AI assistance for restaurant teams, with clear evaluation and human control over consequential changes.

The job

This role explores potential AI features such as menu-drafting assistance, support search and operational summaries. You will connect model capabilities to a well-defined user problem, test whether they help and build the surrounding product experience. The work includes conventional software engineering, evaluation design and cost awareness. Restaurant pricing, publishing and other consequential actions should remain subject to appropriate user review rather than being changed silently by an experimental feature.

What you would be doing

• Build prototypes and production-quality interfaces around carefully scoped AI use cases.
• Design structured inputs and outputs with validation and sensible fallback behaviour.
• Create evaluation datasets and compare usefulness, accuracy, latency and cost.
• Implement retrieval with tenant-aware access controls and source traceability.
• Protect sensitive information through minimisation, redaction and appropriate retention choices.
• Give users clear review and correction steps before publishing generated content.
• Work with product and support teams to identify failure patterns and improve the experience.

What we are looking for

• Strong software engineering skills in TypeScript, Python or both.
• Hands-on experience building an application with model APIs or retrieval components.
• Ability to evaluate model output systematically rather than relying on a few successful examples.
• Understanding of API reliability, validation and secure handling of customer information.
• Product judgement about where automation is useful and where user review is necessary.

Useful, but not essential

• Experience with structured generation, embeddings or retrieval-augmented applications.
• Familiarity with evaluation tooling and model cost monitoring.
• Experience building workflow or support-assistance products.

Tools you might use

• TypeScript
• Python
• Model APIs
• PostgreSQL
• Retrieval tooling
• Evaluation datasets

Your first 90 days

• Days 1–30Choose a bounded user problem and establish a baseline evaluation using representative examples.
• Days 31–60Build a prototype with access controls, output validation and a clear review experience.
• Days 61–90Run a limited evaluation and recommend whether to improve, release or stop the feature based on evidence.

The practical bits

• 3 days per week in the Nottingham office and 2 days working remotely within the UK.
• 37.5 hours per week, Monday–Friday. Core collaboration hours are agreed with the team.
• Occasional planned team or customer visits; approved expenses are reimbursed.

What we offer

• 30 days of annual leave plus public holidays
• A company pension contribution
• A £1,000 annual learning budget for courses, conferences or professional development
• A company laptop and role-appropriate equipment; engineering roles include a development setup
• Flexible start and finish times, agreed with your team around customer and collaboration needs
• Wellbeing support and regular development conversations with your manager
• Reimbursement of approved business travel and customer-visit expenses