May 2024 – Present · 1 yr 10 mos · Mumbai · Remote

• Designed and implemented a Jenkins CI/CD pipeline with a reusable shell script to standardize PySpark deployments across all Spark jobs. This removed the need for job-specific scripts and centralized JAR dependency management, reducing deployment time and saving several hours of manual effort in each release cycle.
• Migrated legacy 100+ Hive tables to Apache Iceberg and built an automated maintenance pipeline. This modernization of the lakehouse architecture significantly reduced HDFS storage usage and removed the need for manual table maintenance.
• Built a Credit Card Customer Feature Store that consolidated customer-level attributes into a single source of truth. This enabled reuse across multiple analytics, campaign, and ML pipelines, reducing duplicate development work for Data Science and Analytics teams.
• Implemented a centralized source control framework integrated with Control-M file watcher jobs. Spark pipelines now trigger automatically only when all upstream data sources are ready, eliminating manual dependency monitoring.
• Optimized existing Spark jobs by tuning configurations and restructuring processing logic, achieving an 80% reduction in processing time and lowering compute costs across the data platform.

Aug 2020 – May 2024 · 3 yrs 9 mos · Mumbai · Remote

• Developed a unified customer risk profiling framework integrating data from 10+ banking products for ~30M customers. This enabled centralized risk visibility across banking channels and was recognized with the Economic Times CIO Award 2024 for Excellence in Technology Implementation – Business Resilience Impact and an internal BIU Star Award.
• Engineered real-time fraud detection pipelines using Kafka and Spark Streaming to process high-velocity transaction data. This enabled instant identification of suspicious transactions and strengthened fraud prevention across digital banking channels.
• Implemented advanced data validation and sanitization processes within data pipelines. This reduced data anomalies by 30% and improved data quality, which increased fraud detection effectiveness by 20%.
• Built a centralized fraud data mart and optimized ETL pipelines. This reduced processing time by ~80% and enabled faster fraud investigations and reporting for analytics and risk teams.
• Collaborated closely with risk, fraud analytics, and compliance teams to design scalable data solutions. These systems helped prevent ₹12+ crore in financial losses and improved customer trust in digital banking platforms.