Tharun Madharaputh Arun

Led ML/NLP initiatives; designed and implemented production-grade ML pipelines for customer churn and retention; containerized models with Docker; integrated with CI/CD using Azure/AWS; built dashboards and monitoring.

Built voter feedback classification models with Python (scikit-learn, NLP tooling) and PySpark; achieved 92% accuracy. Developed employee churn prediction models and designed Apache Spark ETL pipelines on AWS Glue and Redshift; scaled ML workflows with MLflow and integrated Airflow and Azure ML for automated training and monitoring. Created 15+ interactive dashboards (Power BI, Amazon QuickSight) to accelerate executive decision-making. Led enterprise-scale churn prediction and retention strategy pipelines, enabling data-driven decisions in production-like environments.

Cleaned and organized voter feedback data using PostgreSQL and Python (Pandas, NLTK, SpaCy, Hugging Face Transformers), built sentiment classification models, and crafted SQL-based dashboards. Achieved 92% accuracy and reduced manual review effort by 25% during Voter Mood beta phases.