Hi, I'm Hrushikesh

Business Problem

Gesa's finance and marketing teams lacked a shared, reliable view of loan performance, member growth, and operational efficiency because of manual, Excel-based reporting that delayed insights and introduced data inconsistencies.

Approach & Key Actions

• Built interactive Power BI dashboards tracking loan performance, member growth, and branch KPIs, improving transparency for leadership.
• Designed SQL data models and automated pipelines consolidating member, deposit, and loan data, cutting manual reporting effort by 35%.
• Partnered with marketing and product teams to measure campaign effectiveness, identifying segments that lifted offer conversion by 12%.
• Optimized DAX logic and Power Query transformations, standardizing data definitions and boosting dashboard reliability by 20%.
• Delivered quarterly financial and risk reports, automating validation and reconciliation workflows that achieved 100% compliance with reporting standards.

Impact & Results

• Enhanced KPI visibility across finance and branch operations.
• Reduced report preparation and reconciliation time by over 35%.
• Increased dashboard adoption and trust across business units.

Business Problem

The client was experiencing high churn that reduced revenue and raised acquisition costs, without a reliable way to surface at-risk customers early.

Approach & Key Actions

• Analyzed over 1 million banking transactions and behavioral signals to reveal churn triggers.
• Engineered predictive features, highlighting inactivity periods and late payments as strongest indicators.
• Benchmarked Logistic Regression, Random Forest, and XGBoost models, selecting XGBoost (AUC 0.89).
• Published Tableau dashboards to deliver churn scores across 50,000+ accounts.
• Deployed a real-time scoring pipeline with AWS SageMaker orchestrated in Airflow.

Impact & Results

• Reduced overall churn risk by ~15%.
• Improved retention campaign targeting efficiency by ~20%.
• Lowered retention spend by ~12%.

Business Problem

Distributors and retailers were missing critical alerts because SMS and app notifications were unreliable, creating poor user experiences and risking churn.

Approach & Key Actions

• Aggregated 2M+ daily SMS and in-app events into a unified performance pipeline.
• Performed root-cause analysis on delivery logs to isolate operational bottlenecks.
• Automated NLP-based alert categorization to lift reporting accuracy by 28%.
• Ran A/B tests on send times and copy to maximize interaction.
• Built Power BI dashboards for real-time engagement visibility.

Impact & Results

• Increased retailer engagement with notifications by ~19%.
• Reduced failed notifications by ~23%.
• Improved reliability between FMCG distributors and their retail network.