Aug 2024 – Present · 1 yr 9 mos · Bengaluru, Karnataka, India · Hybrid

• Core contributor to the architectural evolution and operational resilience of Scribe, Wayfair’s internal event-streaming platform that acts as the company’s central nervous system—processing 15B+ primary events/day across 350+ event types to power real-time analytics and marketing attribution.
• Architected a high-scale event observability framework spanning a 3-stage pipeline (S1 → S2 → S3), tracking 45B+ records/day via Google Pub/Sub. Built reactive, non-blocking consumers using Project Reactor, enabling efficient concurrency handling and predictable latency under extreme throughput.
• Designed and implemented asynchronous, event-driven infrastructure services based on Reactive Streams, supporting ingestion, validation, and schema resolution without thread blocking—maximizing throughput during bursty workloads while enforcing backpressure.
• Improved system resilience and availability by eliminating a critical SPOF in Scribe 2.0 through a redesigned Schema Registry persistence layer backed by Aerospike, ensuring high availability across 5 global data centers and supporting non-blocking access patterns.
• Led modernization of legacy EIM Kafka pipelines to Google Cloud Dataflow, introducing a Producer-Owned governance model that improved cost attribution (Chargeback/Showback), clarified incident ownership, and simplified operational debugging.
• Guaranteed 100% uptime during peak retail events (Way Day, Black Friday) through region-wide capacity planning combined with reactive backpressure mechanisms to safely absorb traffic spikes without cascading failures.
• Engineered fine-grained, reactive data observability across the Scribe V2 backbone to detect and remediate previously “silent” 504 drops, reducing investigation cycles from weeks to hours. Additionally, developed automated cost-extraction pipelines to support leadership’s Build vs. Buy analysis for Confluent Cloud.