Agraja Jahirdar

Led Azure Compute-side design and delivery of Disk IO Pacing, a fleet-scale reliability feature protecting storage stamps during high ingress, actively pacing 5K disks per day across 500 clusters in production. Drove improvements in time to failure and helped eliminate backend disk failures once pacing was engaged, enabling fleet-wide stability. Designed high-capacity, high-IOPS load tests to convert near-outage conditions into bounded, recoverable behavior. Owned production telemetry and reporting, providing real-time fleet-wide visibility to identify regional hotspots and stability trends. Partnered on secure code signing and CI/CD regression checks, enforcing fleet-wide security standards. Mentored new hires through the Onboard in Buddy program and contributed to on-call support.

Designed and implemented core features for Tier-1 microservice handling high-volume traffic with sub-100 ms latency using Java Spring Boot and Kafka. Led multi-geo Kafka migration, improving message throughput by 40%. Optimized Tomcat thread pool configuration and JVM parameters, increasing concurrent request capacity via heap dumps analysis and GC tuning. Built multi-configuration Jenkins CI/CD pipelines supporting parallel deployments across regions. Conducted unit, integration, system, load, and drill testing to ensure features delivered via platform components and services were highly scalable, available, and reliable. Mentored junior developers and performed code reviews.

Built a full-stack serverless web app using Angular 2, MySQL, AWS Lambda, and API Gateway. Implemented core features and contributed to the overall architecture, delivering scalable endpoints and robust data handling.