The mission
Build a state-of-the-art ML platform and the discipline around it.
- Models with a price tag. Every model has a business case and a measurable monetary outcome.
- Credit and decisioning models built with rigorous validation, champion/challenger testing and explainability.
- A production ML platform. Serving, monitoring, reproducibility and retraining are engineered, not improvised.
- Responsible AI, built in. Model risk, bias and explainability checks, with an independent sign-off gate before anything reaches production.
- Agent-first systems. Agents are production components with orchestration, guardrails, evals and observability.
- ◆ Models on governed data. You build on a point-in-time-correct feature store, not around it.
This is a business function. Every model carries monetary value, and you will run the function that way: compute budget, headcount and return on investment.
What you'll own
The agentic standard for ML. You run agentic model iteration and agentic model monitoring and observability, bound to the feature store. You own the reliability, cost and latency of agent loops and graphs in production, and you hire against the bar.
- The ML & MLOps team, from your first hire onward.
- Model-development standards and the validation methodology that stands up to model-risk and regulatory scrutiny.
- The ML platform behind decisioning services.
- A clear ownership line between feature production (data engineering) and model consumption, set together with the Head of Data Engineering and the Director.
You are
- A leader who loves data and loves building systems around it.
- Hands-on when needed, especially with AI on board. You understand the model, the pipeline and the serving layer.
- Experienced across the full ML lifecycle: development, validation, deployment, monitoring and retraining.
- Experienced in architecting agentic systems for ML: orchestration, guardrails, evals in CI, human-in-the-loop checkpoints, agent observability.
- Experienced in building and leading ML teams in regulated settings.
- Experienced in credit-scoring or underwriting modelling, or comparable high-stakes ML.
- Skilled in model-risk management and responsible-AI governance.
- Experienced in building and leading a team from zero.
- Fluent in English (B2+).
Bonus
- CCD2 and consumer-credit regulation · DORA/ICT risk · IFRS 9 implications for model outputs · fraud-detection ML · Databricks/Spark.
What you will work with
The ML platform (serving, monitoring, and retraining), the feature store, the agent evaluation and observability layer, and cross-functional partners in Data Engineering and Credit Risk.
Why this one
- Seat at the table on a core leadership team.
- Build it right the first time. No legacy ML estate.
- Models that matter. Your work decides real money, not a dashboard.
- Real pace. A lean, AI-native organisation.