Sr. Data Scientist - Credit Risk (IITs/NITs/BITs)

About The Opportunity

• Lead end-to-end development of credit risk models (PD/LGD/EAD): data sourcing, feature engineering, model building, benchmarking and documentation.
• Design and implement robust model validation, back-testing and performance monitoring frameworks aligned to model-risk governance and regulatory expectations.
• Productionize ML models: build scalable pipelines (PySpark/Python), collaborate with engineering to deploy models and automate scoring in lending platforms.
• Apply advanced machine learning (tree-based ensembles, stacking, calibrated probabilities) to improve discrimination and calibration for scorecards and risk ranking.
• Partner with stakeholders (product, analytics, compliance) to translate business requirements into model objectives and measurable KPIs.
• Mentor junior analysts, define modelling best-practices and contribute to model risk documentation and audit-ready deliverables.
  

Skills & Qualifications

Must-Have

• Python
• SQL
• PySpark
• scikit-learn
• XGBoost
• PD/LGD/EAD modelling
  

Preferred

• Experience with model monitoring and MLOps tooling (CI/CD for models, Airflow)
• Knowledge of regulatory frameworks (Basel II/III, IRB) and model governance
• Familiarity with cloud platforms (AWS/GCP/Azure) for data and model deployment
  

Additional Qualifications

• Proven track record in building credit risk models for retail, SME or corporate lending portfolios.
• Strong analytical rigor with experience in statistical validation, uplift testing and calibration techniques.
• Preferred: educational background from IITs/NITs/BITs or equivalent technical institutions.
• This is an on-site position based in India; ability to work closely with cross-functional teams and attend client meetings is required.
  

• Hands-on ownership of high-impact credit-risk projects and visible client exposure.
• Mentorship-driven environment with focus on technical excellence and model governance maturity.
• Opportunities to work on regulatory-compliant implementations and scale models into production.