Job Title: Data Scientist
Experience: 3 to 6 years of relevant data science experience
About the Job:
We are looking for Data Scientists who can independently own a modelling workstream and deliver practical improvements to Tonik's lending decisions. Depending on allocation, you will work on cash or installment-loan underwriting, behavioural and offer scores, collections, payroll-backed lending, fraud, offer propensity or income prediction.
You will develop and validate models, investigate performance changes and work with MLE to bring approved models into production. The team is moving toward risk-based refreshes, generally every three to six months for established scorecards, supported by timely monitoring and stronger development discipline.
Job Responsibilities:
• Own assigned modelling workstreams from business problem definition and data preparation through feature engineering, model development, validation and production handover. Agree the target, population, success measures and delivery plan with stakeholders.
• Build features from application, demographic,
bureau, device, event, transaction, repayment and alternative data. Use Python and SQL to create reproducible datasets and verify joins, missing values, timestamps and customer or loan identifiers.
• Develop and compare scorecards and machine learning models using appropriate statistical and gradient boosting methods. Evaluate component scores, stacked models and segment-specific challengers against credible baselines.
• Design temporal and out-of-time validation, prevent target leakage, and ensure repayment outcomes have matured. Identify distortions caused by restructuring, account freezes, approval policy changes or changing collection strategies.
• Evaluate discrimination, calibration, stability and segment performance using Gini, KS, PSI, CSI and relevant business measures. For the assigned use case, relate model performance to credit losses, recoveries, attrition, offer uptake or predicti
📌 Data Scientist (Chennai)
🏢 Tonik
📍 Chennai