SANDHYA S

AutoInsights:

• Developed a production-ready forecasting model using a Random Forest Regressor that improved sales forecast accuracy to 87%.
• Engineered a feature combination analysis module using a Random Forest Regressor to identify and rank high- and low-impact categorical feature groups (e.g., via feature importance and normalized impact scoring).
• Engineered a multi-cloud generative AI feature, leveraging AWS and Vertex AI, to generate custom insights and charts.
• Designed and implemented a scalable data processing architecture that ingests diverse data sources (DB tables, Excel files) for insight generation.
• Productionized the product using AWS EC2, creating a RESTful API, and configuring a CI/CD pipeline with Nginx, Flask, and Gunicorn.

AppsIQ:

• Built and deployed an agentic framework using AWS to automate CRM analytics workflows, reducing insight generation time to under 10 seconds.
• Developed a knowledge base to document API structures, which enhanced scalability and responsiveness by 80%.

Churn Prediction:

• Developed a Random Forest Classification model to predict customer churn risk, achieving 85% accuracy in classifying active period days.
• Implemented cross-validation to ensure model robustness and generalizability.

NielsenIQ:

• Automated flat-file analysis generates metric-based outputs, reducing report generation time to under 10 seconds.
• Improved reporting efficiency by 60% through automation, while implementing data quality checks to ensure accuracy.

• Automated UI regression testing for a telematics application, using Python's BDD framework.

• Developed an anomaly detection system using EIF on AWS logs processed via Elasticsearch, improving RCA.