About the role
This is one of the most important roles we are opening. You will work on the core of where Taily is heading: automatically normalising supplier data, and a reconciliation engine that guards that everything adds up. This is not standard machine learning work. It is about reliability: suppliers deliver data in dozens of formats and dialects, and it has to line up faultlessly with what is already in the customer's systems. Our starting point is strict: AI proposes, a deterministic system executes, and a human checks.
What you will do
- Help build AI-driven normalisation of supplier data (colours, sizes, categories) into the customer's taxonomy
- Work on the reconciliation and state-estimation layer: deciding what belongs where, and spotting deviations before they go wrong
- Embed models in a system with deterministic checks and a human review step, instead of trusting output blindly
- Build our cross-customer mapping corpus as structured, searchable data, so the system improves the more it sees
- Back everything with tests and evaluation: demonstrably correct beats fast
Technical stack
- Python for the ML and normalisation work
- Integration with our Ruby on Rails application
- PostgreSQL, Redis, Sidekiq
- DuckDB for analysis
- AWS
- GitHub
- A background in robotics, state estimation or another field where uncertainty and correctness are central is a strong plus
What we offer
A role with real technical depth and direct impact on the product.
- Work on a problem with no off-the-shelf solution
- A team that takes correctness seriously and does not ship things lightly
- Flexible hours, hybrid working possible
- A modern office in the centre of Den Bosch
- A market-rate salary
- Share options, to be discussed
- On the right match, room to help build the company, remuneration included