Contract Credit Risk Modeller

Contract Credit Risk Modeller

• £560-£640 per day
• Inside IR35
• Fully remote
• Three-month contract

Harnham is partnering with a leading financial data and analytics organisation to recruit a hands-on Credit Risk Modeller for an initial three-month engagement.

You will join its UK Product Analytics and Innovation team, supporting the rapid development of a new credit-risk product.

You will independently build a point-of-application credit risk scorecard from raw data through to a validated, production-ready MVP.

Your responsibilities will include:

• Ingesting, joining and cleaning complex customer and credit datasets
• Constructing development samples and defining observation and performance windows
• Establishing appropriate good/bad definitions
• Engineering characteristics and treating missing and special values
• Performing monotonic binning, Weight of Evidence and Information Value analysis
• Developing an interpretable logistic-regression scorecard
• Completing feature selection, model tuning and points-based score scaling
• Validating the model using measures such as Gini, KS, AUC, PSI and out-of-time testing
• Producing clear technical documentation and supporting production implementation

The successful candidate will have:

• Personally built and deployed at least two end-to-end consumer or commercial credit scorecards
• Advanced hands-on Python experience within credit-risk modelling
• Strong SQL skills and the ability to prepare complex modelling datasets independently
• Practical experience with WoE, IV, binning, logistic regression and score scaling
• Experience defining modelling samples, performance windows and credit outcomes
• A track record of implementing scorecards within lending, underwriting, collections or credit decisioning
• The ability to write custom transformations and debug Python logic independently
• Experience delivering models within regulated financial-services environments
• Availability to begin immediately or at short notice

Experience with credit-bureau, SME, commercial, Companies House or Open Banking data would be beneficial. GCP and BigQuery experience is also desirable, although other cloud platforms will be considered.

Candidates should be prepared to discuss a previous scorecard build in granular detail during the interview, including feature choices, binning decisions, model trade-offs, validation results and their individual coding contribution.