About the role
Our positioning is literally "proper science underneath, AI on top", and this role is the underneath. The pricing engine learns how customers respond to price moves, the forecast warns shops weeks before a dip, and the budget allocator moves ad money to what earns after cost price and shipping. Every one of those models is yours to improve, validate and extend.
You work with real, messy, wonderful data from more than ten pilot shops across industries, from small specialists to department stores with thousands of products. Your models do not end in a notebook: they reprice real products the same night, inside guardrails, with every decision logged. Honest evaluation matters more here than fashionable architecture.
What you will do
- Own the elasticity, forecasting and budget allocation models end to end
- Design honest evaluation: baselines, confidence bands and error reporting a shop owner can trust
- Ship models into production with the engineering side, including monitoring and rollback
- Ground Sherpa, the AI layer shops talk to, in numbers that are actually correct
- Say clearly when a simpler model wins, and prove it
What you bring
- Strong statistics and classic ML fundamentals: regression, time series, causal thinking
- Production experience: your models have run somewhere real, with consequences
- Fluent Python or equivalent, comfortable SQL, respect for data quality
- The instinct to distrust your own results until they survive a hostile look
- The right to work in the UK; we cannot support visa sponsorship
Nice to have
- Pricing, demand or marketing mix modelling experience
- Retail or eCommerce data experience
- Published work, talks or a blog that shows how you think
Compensation
From £80,000, very competitive and rising with experience
Hybrid across the UK: work from home plus WeWork locations in London and worldwide. You must already have the right to work in the UK; we cannot support visa sponsorship.
Please note: we do not accept applications from freelancers, the self employed or agencies.
Apply
Tell us who you are.
No cover letter theatre. Your name, a way to reach you, a link to what you have done, and a few honest lines. A human reads every application.