Abhinandan Samal

• Architected real-time data pipelines using Confluent Cloud (Kafka, KSQLDB), building 50+ production stream-processing objects for complex JSON/AVRO event ingestion. • Designed a custom Kafka deserialization framework, reducing error-processing time by 30% and improving downstream analytics and ML reliability. • Automated cloud infrastructure deployment using Terraform and Azure DevOps CI/CD, ensuring scalable, compliant, and auditable data platforms.

• Built and deployed end-to-end ML pipelines for customer churn prediction using LightGBM, CatBoost, Random Forest, achieving >90% recall in production. • Implemented automated retraining workflows on AWS SageMaker, reducing manual maintenance and improving model stability. • Developed anomaly detection and recommendation systems on 13M+ records, increasing cross-selling by 15% and upselling by 9%.

• Designed and maintained batch and streaming ETL pipelines using PySpark, Python, and SQL. • Delivered 4 new recommendation use cases by building reusable, scalable data-processing components. • Led a PySpark migration POC that significantly improved batch processing performance and system throughput.

• Automated operational reporting using UNIX shell scripting and PL/SQL, reducing manual effort and improving reporting accuracy. • Supported database migrations and built dashboards using Power BI for management reporting.