Data Scientist

Job location: Lagos, Nigeria

Experience: 3-6 years

Monthly salary: USD 1,500–3,000

Local allowance: USD 400 equivalent in local currency

Accommodation provided by the company

Company-provided car

Immigration expenses covered by the company

One-time travel expense to India for a 30-calendar-day leave

Role summary:

We are looking for a Data Scientist to build and run the models like: credit risk scoring,

propensity and growth models, and fraud and anomaly detection, embedded directly into our lending and wallet products. The candidate will work alongside the Head of Data, Financial Services to turn platform data from distributors and retailers into data-led decisions. The role is not only about building models. It is about proposing the right model for a business problem, deploying it into the product, monitoring it continuously, and helping the business trust and adopt it.

Key responsibilities:

1. Credit risk assessment

• Build, validate and maintain credit risk models: application scorecards, behavioral scores, probability of default (PD) and days-past-due (DPD) roll-rate models

• Use platform data (commerce orders, wallet flows, repayment history) to score distributors and retailers, including thin-file customers with no bureau history

• Recommend credit limits, pricing tiers and cut-offs that balance approval rates against portfolio risk and NPA

• Feed model outputs into early warning signals that flag deteriorating accounts before they default

2. Customer growth and ROI from existing customers

• Build propensity models (likelihood to take a loan, top up, renew, adopt a new FS product) to target acquisition and cross-sell

• Build segmentation and next-best-offer logic to grow customer value and increase ROI from the current base

• Model churn and dormancy risk so the business can act before customers drop off

• Design and read out experiments (A/B and holdout tests) to prove the uplift each model delivers

3. Fraud and anomaly detection

• Build fraud detection models for onboarding, transactions, repayments and agent or field activity

• Build anomaly detection over transaction and wallet behavior to catch unusual patterns, collusion and abuse early

• Combine rules and machine learning, and tune alert thresholds to keep false positives manageable for operations teams

4. Embedding models into products

• Work with product and engineering to deploy models into live decision flows: loan approval, limit setting, offer targeting and fraud checks

• Package models as scores, APIs or batch outputs that product systems can consume reliably

• Make sure each model runs well in production, not just in a notebook

5. Continuous model monitoring

• Set up monitoring for model performance, data drift and population stability (e.g. Gini, KS, PSI, approval and default rates)

• Define retraining triggers and maintain model documentation, versioning and governance

• Report model health to the Head of Data, FS and to risk stakeholders on a regular cadence

6. Data-led automation

• Automate recurring analyses, scoring runs and decision rules so the business moves from manual review to model-driven decisions

• Build reusable feature pipelines and datasets with the data engineering team

7. Stakeholder communication and data adoption

• Explain what a model predicts, how reliable it is, and what it means for the business, in plain language for credit, product, sales and leadership teams

• Translate model results into business impact: approvals, losses avoided, revenue gained, customers acquired

• Champion data adoption by helping teams use model outputs in their daily decisions

Experience and qualifications:

• Bachelor's or Master's degree in Statistics, Mathematics, Economics, Computer Science, Data Science or a related quantitative field

• 3-6 years of hands-on data science experience, with at least 2 years in fintech, digital lending, banking or payments

• Proven experience building credit risk models (scorecards, PD, DPD/roll-rate) on real customer transaction data, ideally for SME, merchant or thin-file lending

• Experience building propensity, segmentation or churn models that drove measurable growth

• Experience with fraud detection or anomaly detection on transaction or payment data

• At least one model taken from idea to production and monitored after launch

• Solid grasp of financial terminology: NPA, DPD, PD/LGD/EAD, portfolio-at-risk, vintage analysis, approval rate, loss rate, CASA, interest and fee income

Technical skills:

Languages and data Strong Python and SQL; PySpark or other big-data tools

Data platform Snowflake or a similar cloud data warehouse

Modelling Logistic regression and scorecards (WoE/IV), gradient boosting (XGBoost, LightGBM),

clustering, survival or time-to-event models

Fraud and anomaly Isolation Forest, autoencoders, graph or network features, rules plus ML hybrids

Explainability SHAP, reason codes, feature importance suitable for credit decisions

MLOps Model deployment (APIs or batch scoring), versioning, drift and stability monitoring (PSI, CSI)

Reporting Power BI or a similar BI tool for model and portfolio dashboards

Communication and behaviors:

• Clear storyteller: explains model accuracy, limits and business impact to non-technical audiences without jargon

• Business-first thinker: starts from the business problem and proposes the model that solves it, not the other way round

• Ownership: takes models through to production and stays accountable for how they perform

• Collaborator: works well with credit, product, engineering, operations and sales teams

Adoption champion: builds trust in data and helps teams change how they make decisions

Nice to have:

• Experience lending to SMEs, merchants, distributors or retailers in Nigeria or other African markets

• Familiarity with CBN regulations, credit bureau data and BVN-based identity data

• Experience with alternative data for credit scoring (wallet, POS or e-commerce transactions)

• Experience building early warning signal systems