Data Scientist, Fraud Risk Analytics (Public Sector Programme) | Contract

Job Responsibilities Review and recalibrate FAM 2.

0 model methodologies and thresholds for prioritised backlog items, including back-testing, impact assessment, analytical specifications, validation approaches, gap analyses and documentation Translate business, policy and scheme requirements into changes to fraud-risk scenarios, indicators, features, thresholds and analytical logic Build and enhance monitoring for model performance (precision, recall), data drift and concept drift, including alert logic, visualisations and deployment support Design performance measures that account for delayed and incomplete investigation outcomes Assess new schemes’ structures, eligibility rules, claims processes and known fraud risks, map relevant patterns against the existing FAM 2.

0 risk taxonomy, and recommend how each scheme can be brought into coverage Maintain clear methodology and change records Communicate findings, trade-offs, dependencies and recommended actions to programme owners, FAM stakeholders, engineers and management Produce reproducible code, analyses and documentation throughout the engagement Job Requirements Degree in Data Science, Statistics, Mathematics, Computer Science, Engineering, Economics or a related quantitative discipline 3 years of hands-on experience in data science, advanced analytics, model validation, fraud/risk analytics or programme analytics Strong hands-on proficiency in Python and SQL for data extraction, transformation and exploratory analysis, with experience building reproducible analytical workflows Hands-on experience developing, validating or enhancing predictive, risk-scoring, fraud-detection or anomaly-detection models, including feature development, back-testing, threshold calibration and performance evaluation (precision, recall and related trade-offs) Experience translating business, policy or operational requirements into analytical specifications, model logic, indicators, rules or measurable risk hypotheses Experience working