RISHABH PANDEY

Roles & Responsibilities

● Model validation and ongoing model review in the home lending domain.

● Mentoring (No direct team) to the team and shaping up for the next level.

Significant Accomplishments

● AVM MODEL VALIDATION: Automated valuation model is a model framework of 16 models. Validation and monitoring all 16 models on a quarterly basis. Validation involves risk guidelines validation, developing challenger models and independent testing.

● RISK RATING EVALUATION: Evaluating and Identifying the risk rating of the AVM models, validation process is depend on the risk rating of the model

Roles & Responsibilities

● Working across key functional business areas to develop and implement analytical projects to optimize operations cost and customer experience.

● Supporting leadership with customer deepening analysis

Significant Accomplishments

● CUSTOMER LIFETIME VALUE: Developed and deployed successfully CLV model. And this model went through the model governance team also before deployment. We have used multi layering of segmentation and discounted cash flow models to calculate CLV for each and every customer.

● WAIT TIME MODEL: Developed a wait time prediction model for all segments using survival analysis to identify the waiting time for the calls in the time of crises.

● REPEAT CALLER MODEL: Formulated and developed predictive repeat caller model using random forest to identify the customer who is going to call us back within a week for the same reason.

● CALL VOLUME FORECASTING: Develop a call volume for casting for different segments. Used LSTM to forecast the volume.

Roles & Responsibilities Significant Accomplishments

• Developed and implemented predictive models for cross-sell, early claims, and retention strategies across key business areas.
• Managed cross-sell campaigns across various channels.
• Created and maintained a data mart providing a comprehensive 360-degree view of the customer lifecycle.
• Predictive Model for Retention: Implemented a logistic regression model in retention call centers to predict policyholders likely to lapse, successfully reducing the lapse rate.
• Policy Reinstatement Analysis: Identified characteristics of policies with high reinstatement probability and developed a predictive model using logistic regression, enhancing customer win-back strategies.
• Fraud Detection: Formulated a fraud score model using random forest to identify early fraudulent claims, aiding compliance teams in early detection.
• Customer Segmentation: Developed segmentation using CART decision trees, enabling call centers to tailor verification calls based on customer profiles.
• 360-Degree Data Mart: Established a comprehensive data mart covering the complete policy lifecycle, streamlining analytics projects and predictive modeling.
• Cross-Sell Prediction: Created a random forest model to identify customers with a high propensity for cross-sell, increasing the policy-per-customer ratio.
• D2C Cross-Sell Campaign: Designed and executed a direct-to-customer cross-sell campaign for unreachable customers, providing call center teams with a shared customer base for effective cross-sell pitching.

Roles & Responsibilities

● Developing, deploying and tracking performance of predictive models for different business areas and problems.

● Alignment of the predictive approach and model to different business end users and teams.

● Solving business problems using statistical techniques and providing end to end automated process flow.

Significant Accomplishments

● DROP-IN USAGE MODEL: Developed and deployed drop in usage predictive model by using logistic regression for different circles and segments. This model can predict which customers will drop their usage in the next 3 weeks.

● USAGE CHURN MODEL: Created predictive model using historical and current data to identify the customer which will show usage churn in next month.

● QUALITY OF ACQUISITION MODEL: Formulated and developed predictive model using logistic regression to capture the behavior of newly acquired customers in the second month from the date of acquisition. This model will help the marketing team to run appropriate campaigns to retain customers.

● CUSTOMER SEGMENTATION: Segmented the whole scored base using cluster approach and share it to the marketing team on which they design segmented promos according to the characteristic of segments.

● ZERO USAGE MODEL: Developed and implemented zero usage algorithms and predictive model by using logistic regression for different circles and segments, this model can predict which customer will be in zero usage base for next seven days.

● GROWTH IN USAGE MODEL: Developed and deployed a model which can predict which customer will show growth in usage in the next month by using CHAID decision.

● MNP model: Designed and developed a predictive scoring model by using random forest (Machine learning technique) to optimize call center’s retention operation, which directly impacts the MNP conversions.

● AUTOMATING BUSINESS REQUIREMENTS: Solving different business problems and providing end to end automated process flow. Below are the few accomplished process flow

o Developed SAS process flow which can bifurcate the performance file of the sales department and send the performance of employees to their respective managers.

o Developed an automated process of capturing spike, level shift and trend in daily level data through SAS macro. This will help to identify uneven behavior in data.

Designed an automated process to track the performances of all predictive models. This will help to track model performance and identify which model needs to be redeveloped.