Sep 2022 – Present · 3 yrs 9 mos

• Led architecture of Azure’s real-time billing engine ingesting usage events, matching against millions of price points, and
• applying discounts to compute accurate charges at scale; processing 500M+ events/hour with P99 latency <5 min, powering
• billions in annual revenue; built on Event Hubs, Cosmos DB, Delta Lake
• Designed and implemented re-rating system handling mid-cycle price changes using Synapse and change tracking to identify
• affected records, recomputing charges across 100M+ records within 1 hour while meeting invoice finalization SLAs
• Architected and implemented discount allocation for a multi-million-dollar savings plan business; designed search and sorting
• algorithms matching usage against optimal discounts across varying rates, achieving 5x throughput improvement
• Built pre-classification layer determining usage-to-discount eligibility upfront, reducing allocation computation by ∼90%
• Engineered active-active failover with checkpointing for zero data loss; resolved duplicate events via unique ID deduplication,
• achieving zero duplicate billing errors across the platform
• Engineered end-to-end platform for reservation application, utilization reporting, and revenue recognition; built catalog sync
• using Azure Functions with TTL-based caching across thousands of SKUs and regions
• Architected cross-subscription reservation pooling, enabling enterprises to share compute reservations across organizational units,
• driving major adoption and revenue growth
• Led security remediation addressing 20+ compliance gaps across 100+ resources; designed network isolation using Private
• Endpoints and Network Security Perimeters with Managed Identity for credential-free authentication

Jul 2018 – Sep 2022 · 4 yrs 2 mos

• Led architecture of recommendation systems across cloud product lines (VMs, SQL, Storage, RedHat, SUSE, Synapse); analyzed
• hourly usage patterns, dynamically adjusted forecasts, and applied feature engineering, cost-benefit modeling, and ranking
• algorithms
• Optimized Spark pipeline via dynamic allocation and scheduling tuning, achieving ∼58% runtime reduction (13h to 5.5h);
• generalized into pluggable framework for rapid product line onboarding

Nov 2016 – Jul 2018 · 1 yr 8 mos

• Designed and implemented backend enabling first-ever transactable SaaS purchases on Microsoft’s cloud marketplace (flat-rate,
• per-seat, free trial); integrated purchase, fulfillment, licensing, and cancellation workflows
• Built search, indexing, and retrieval systems (product search, orders, invoicing, inventory) using Cosmos DB, Azure Search, and
• event-driven microservices (Service Bus)

May 2015 – Jul 2015 · 2 mos · Greater Hyderabad Area

• Enhancing Accessibility of PDFs in the Microsoft Edge Browser.
• To use the metadata information in the PDF/UA for assisting the PDF accessing
• applications.
• This information is used by the applications for easy navigation across the PDF.