Senior Data Scientist

Overview A fast-growing fintech is hiring a Senior Data Scientist on a permanent, full-time basis in London, with at least three days a week in the office (Mondays are a fixed in-office day for the whole team). The salary is £94,589 to £111,502 depending on experience, with stock options of £80,479 to £105,003 on top. Benefits include 35 days holiday, private healthcare with Bupa, up to 12 weeks enhanced parental leave, a £750 annual learning allowance, and a cycle-to-work scheme. This role exists because the business is still building its predictive modelling capability, not inheriting a finished one. You will own the full modelling lifecycle across credit acquisition, portfolio management, collections, fraud and rewards: from initial hypothesis and feature engineering through to production deployment, monitoring and iteration. Your outputs feed directly into a credit strategist's decision-making, so clarity and usability matter as much as model performance. You will report to the CRO and work shoulder-to-shoulder with the teams making live commercial decisions. If you want to build something that genuinely shapes how a consumer credit business runs, rather than polish models in a central team, this is the role. Key Responsibilities

• Build and continuously improve acquisition models using bureau, application and alternative data to assess applicant risk, affordability and expected customer value
• Develop behavioural models for portfolio management, helping the credit strategy team improve credit profitability
• Build collections segmentation models that predict member responses to different interventions and focus effort where it has the greatest impact
• Develop and improve fraud detection models, identifying suspicious behaviour earlier and reducing friction for genuine members
• Work with AI engineers to use member data to surface the right reward at the right time
• Own the full modelling lifecycle: data exploration, feature engineering, training, validation, implementation, and ongoing performance monitoring
• Get models into production yourself, not hand them off to engineers
• Identify additional internal signals, bureau attributes and external data sources that could materially improve prediction
• Build robust model monitoring and validation so the team understands calibration, discrimination and drift after deployment
• Explain model outputs and methodology clearly to the CRO and committee members without relying on technical language
• Help establish the modelling standards, tooling and experimentation practices that will define how data science works at the business Requirements Must-haves
• Predictive models shipped to production and driving real decisions
• End-to-end ownership: feature engineering through post-deployment monitoring
• Python fluency and strong SQL
• Deep understanding of model evaluation beyond headline accuracy: calibration, stability, drift
• Comfortable with ambiguous problem statements and messy, unstructured data
• Able to communicate model outputs and methodology clearly to non-technical stakeholders
• Commercial mindset: optimises for decision quality, not modelling sophistication Nice-to-haves
• Experience in consumer lending, credit cards or fintech
• Hands-on work with bureau data and behavioural credit models
• Background in fraud detection or collections modelling
• Experience with transaction data or customer recommendation systems What Success Looks Like
• Acquisition, collections and fraud models are live in production within the first six to twelve months, with measurable impact on decisions
• The credit strategist has clear, usable model outputs they can act on with confidence
• Models are monitored consistently and iterated on when performance drifts
• The CRO and committee members can follow and interrogate model logic without needing a technical interpreter
• Modelling standards, tooling and development practices are meaningfully more mature than when you joined Team and Culture
• Small, senior team where you work directly with the people making commercial decisions rather than through layers of stakeholders
• Office-first culture, with genuine emphasis on proximity between data science and the teams using its outputs
• The CRO owns both risk and analytics, so data science sits at the centre of the business rather than at the edge of it
• Builder's culture: speed and real-world impact are valued over theoretical perfection Challenges
• The modelling stack is still being built, so you will need to create structure and standards as well as deliver models
• The remit is broad: acquisition, portfolio management, collections, fraud and rewards all sit within scope for one person
• Working with ambiguous problem statements and incomplete data is the norm, not the exception
• Stakeholder communication is a genuine part of the job: explaining complex modelling decisions to a CRO and committee requires consistent clarity and confidence