Data Scientist

Alba Partners is recruiting two experienced Data Scientists to join a Leading Tier 1 Financial Services client working alongside an existing delivery team to design, build, validate and productionise AI and machine learning models that support supervisory and operational work.

The client is looking for scientists who have already taken models from development through to live, production-grade deployment, and who can turn supervisory and operational problems into deployed analytical solutions that hold up in a regulated environment. You will work closely with other data scientists and business teams to understand the problem, build the model, validate it properly, and get it into production alongside the documentation and services that support it.

What you will be doing

Designing and developing AI and ML-based solutions against real supervisory and operational problems. Working with other data scientists to build and deploy production-level solutions, not just proof of concept. Troubleshooting and debugging code across the model lifecycle. Working with business and technical teams to translate ambiguous problems into deployable analytics.

Solid, current hands-on experience with Python for data science: pandas, NumPy and scikit-learn for data wrangling, modelling and feature engineering. Strong SQL for querying structured data sources. Genuine model development and validation experience across classification, unsupervised learning such as outlier detection, and ranking models. Real machine learning deployment experience: containerised deployment using tools such as Podman, SageMaker or DSW pipelines. This needs to be production experience, not just notebooks. Version control with Git, working reproducibly and collaboratively. Time-series analysis, including assessing risk trends across financial years. Exploratory data analysis, spotting early signals and risk clusters in complex data.

Desirable

Rank aggregation or ensemble techniques such as Robust Rank Fusion (RRF). Model explainability tooling such as SHAP or LIME. Model monitoring and drift detection. RegTech, financial crime or data-led supervision project experience, this is a strong plus given the nature of the work. Record linkage and/or network analytics. Graph query languages such as Gremlin or Cypher, and graph databases such as Neptune or Neo4j, including graph visualisation.

Who this suits

This role suits a data scientist who has genuinely shipped models into production, not just built them, and who is comfortable operating in a regulated, public-sector environment where documentation, validation and defensibility matter as much as model performance. Experience in financial services, regulation or financial crime is a strong advantage but not a prerequisite if your production ML track record is strong.