• Built a petabyte-scale data lakehouse platform on GCP using Apache Spark, Iceberg, and BigLake, handling hundreds of millions of records daily across thousands of datasets. Serves data scientists, analysts, and product teams across the org.

• Architected a high-throughput stream processing platform on Apache Flink for low-latency workloads at scale. Four teams now run their own pipelines end-to-end with no platform dependency.

• Designed a self-service dataset ingestion platform that cut onboarding from weeks to hours, with schema evolution, deduplication, and late-arriving data handled out of the box.

• Tackled hard distributed systems problems at scale including spatial indexing, stream deduplication, and partition-aware data reconciliation.

• Shipped two Generative AI tools into production, a conversational onboarding agent and an LLM-powered debugging tool for on-call engineers.

• Own the technical roadmap and partner with product, leadership, and cross-functional stakeholders to align platform direction with business priorities.

• Lead design reviews, mentor engineers, and work closely with leadership on hiring and team growth.

• Designed and built a multi-tenant data management platform in Java, letting customers create, query, and maintain datasets in minutes rather than days. Implemented optimistic concurrency to handle concurrent access across thousands of tenants with strict isolation.

• Owned design and delivery of a disaster recovery system with automated discrepancy detection and reconciliation across distributed replicas within SLA.

• Drove decomposition of a monolithic Java application into independently deployable microservices. Defined service boundaries, API contracts, and interface specs that became the foundation for years of service extraction work.

• Delivered lazy loading and caching improvements that cut memory footprint by 37% and lifted throughput by 12%. Rebuilt test infrastructure for TDD, cutting build times by 20%.

• Ran design reviews and mentored engineers across the team.

• Built distributed event collection infrastructure for real-time clickstream processing using Kafka and Redshift. Developed low-latency REST services for near real-time scorecards and statistics.

• Ran A/B testing initiatives across user engagement strategies, driving a 20% increase in user involvement

• Designed and built RESTful backend services in Java on Hadoop/HBase and MongoDB for content management at scale. Cut procurement cycle times by 15%.

• Reworked backend components into a more extensible architecture, reducing dev time for new features by 30%.

• Designed a service-oriented architecture for complex, high-volume B2B procurement transaction flows.

• Re-engineered inventory management and order processing from a synchronous monolith into a distributed async system, improving throughput by 35% and eliminating single points of failure.

• Built a multi-marketplace recommendation system for matching buyers with best available deals across platforms.