We are looking for an experienced, hands-on Credit Risk, Sr. Staff Data Scientist who is comfortable working with large data sets, coding in SQL and Python and gaining insights from the data and translating the results into actionable insights for business stakeholders. In this role, you will maintain and enhance our credit risk models/policies to monitor the portfolio and gain insights. You will also build and monitor credit risk models with an eye on loss forecasting and communicate the results to different teams such as Capital Market and Marketing. The candidate should have a passion for streamlining processes and building tools which can monitor models/portfolio effectively. You will be a key contributor to our risk management processes.
Key Responsibilities
- Building, maintaining and enhancing credit risk models for lending portfolios.
- Extract, clean and manipulate large data sets using SQL and Python; build pipelines and analytics to perform model and portfolio monitoring.
- Perform exploratory data analysis (EDA) to identify portfolio trends, drivers of loss performance (vintage, credit bands, borrower attributes, macro factors) and provide insight into model deviations.
- Maintain forecast deliverables: monthly/quarterly loss forecasts by vintage and segment, stress and scenario analyses, sensitivity testing.
- Provide commentary and insights to business stakeholders on credit policy assumptions, model health, and emerging portfolio risks.
- Automate reporting, dashboards and pipelines to streamline model monitoring and improve efficiency and accuracy.
- Document model methodologies, assumptions, data sources and results in clear, audit-ready format consistent with risk governance requirements.
- Participate in governance and review of credit model methodology, model validation support and liaise with external auditors or regulators where needed.
- Continuously identify opportunities to improve credit decisioning accuracy, data infrastructure, modeling techniques, and integrate advanced statistical or machine-learning techniques as appropriate.