Senior Data Scientist Predictive Analytics (Pune)

Job Summary

We are seeking a highly skilled Senior Data Scientist with expertise in Predictive Analytics, Machine Learning, Explainable AI, and Azure AI Services to develop next-generation predictive models and AI-powered analytical solutions. The ideal candidate will have solid experience in healthcare, insurance, or benefits analytics, with the ability to transform large-scale enterprise data into actionable business insights.

This role will focus on building predictive utilization models, designing explainable AI solutions, and developing conversational AI agents using Azure technologies to enable business users to interact with enterprise data through natural language.

Key Responsibilities

• Design, develop, and validate predictive models for Baseline Utilization Forecasting using historical claims, member demographics, plan design, diagnosis, and procedure data.
• Build attribution and explainability frameworks to identify key drivers behind differences between expected and actual utilization.
• Perform advanced feature engineering on structured healthcare and insurance datasets stored in SQL Data Warehouse, Azure Data Fabric, OneLake, and Databricks.
• Develop and optimize statistical, machine learning, and time-series forecasting models to improve prediction accuracy.
• Build and validate actuarial-style frequency and severity models using:
• Poisson Regression
• Negative Binomial Models
• Gamma Regression
• Tweedie Models
• Conduct model validation, back-testing, and performance monitoring against historical outcomes and existing forecasting methods.
• Implement Explainable AI (XAI) techniques such as SHAP values, feature importance, and driver attribution to provide transparent model insights for business stakeholders.
• Collaborate closely with Data Engineering teams to ensure high-quality, production-ready datasets and semantic consistency across analytical platforms.
• Work extensively within Azure Data Fabric, including OneLake, Lakehouse, Fabric Da .