At Amazon, I led the design and delivery of a serverless S3 partition-warming system across 26 AWS regions, eliminating a 503-throttling failure mode that could trigger cluster-manager re-election loops and node outages.
I also contributed resiliency fixes to open-source OpenSearch, adding repository I/O timeouts and exponential-backoff retries so stuck snapshot operations could recover automatically.
At Amazon, I built storage observability for AWS OpenSearch, instrumenting EFS metrics and adding alerting across 100+ file systems. Earlier, as a Software Development Intern at Tata Power, I developed a utility-data dashboard with data-cleansing and outlier-detection workflows.

