Kissht - Collection Analytics & Modeling Role

Description:

Job Overview:

- The role involves developing and optimizing scorecards, early warning models, and segmentation frameworks to improve recovery rates and reduce delinquencies.

- You will collaborate with risk, operations, and product teams to design data-backed strategies, run A/B tests, and track key collection KPIs.

- Strong skills in SQL, Python, and BI tools are essential, along with a solid understanding of collection workflows and analytics-driven decision-making.

Key Responsibilities:

- Develop, monitor, and enhance collection scorecards and early warning models to predict delinquency and optimize recovery efforts.

- Perform vintage and roll-rate analyses, flow rate tracking, and lag-based recovery modelling.

- Design segmentation strategies for bucket-wise, geography-wise, and product-wise collections.

- Partner with operations to run championchallenger strategies and A/B tests to optimize field and digital collections.

- Build dashboards to track collection KPIs, including DPD movement, cure rates, and agency performance.

- Work closely with credit policy and risk to ensure collection strategies align with risk appetite and loss forecasts.

- Collaborate with data engineering and product teams to enable real-time portfolio insights and collector performance analytics.

Required Skills:

- Strong proficiency in SQL, Python, and Excel for data analysis and automation.

- Experience with statistical modelling, logistic regression, and machine learning techniques for risk prediction.

- Hands-on with BI tools (Power BI / Tableau / Looker) for visual analytics.

- Understanding of collection operations workflows, dialler strategies, and agency management KPIs.

- Ability to translate insights into actionable business recommendations.

Preferred Qualifications:

- 4-10 years of experience in collections or risk analytics within consumer lending.

- Exposure to digital collection platforms, tele-calling optimization, or field force productivity models.

- Educational background in Statistics, Mathematics, Economics, Engineering, or related field.