Data Engineer
VISA
Apr 2025 – Present
Bangalore, India
- Improved NER model precision, recall, and F1-score by 30% via training pipeline optimization and automated evaluation workflows.
- Reduced large-scale NER inference latency by 35% using GPU-backed processing across 10% of Visa’s enterprise dataset
- Engineered a secure, scalable backend data service processing billions of records via streaming and batch integrations, improving throughput by 50%
- Built scalable backend services integrating external systems with secure (SSL/TLS) communication
- Integrated external Natural Language Query services and improved prompt logic to reduce hallucinations by 50%, resulting in consistently accurate outputs
- Implemented JDBC-based data access via Trino within Spring Boot APIs to retrieve millions of records efficiently, securing query execution using parameterized SQL to prevent injection vulnerabilities.
- Implemented resilient retry mechanisms using Spring Retry to handle transient network failures and ensure reliable data processing without loss
JavaSpring JDBCJPA / HibernateSQLNERDistributed SystemsJenkins CI/CD
