I build backend systems
that hold up under pressure.

• Architected a data archival system that unblocked sales for 10+ enterprise customers and ensured data sovereignty compliance across 30+ organizations.
• Shipped a Survey PII processing engine that directly secured 5 enterprise contracts.
• Built a feature-flag and plan-control system enabling granular tier-based access and self-serve monetization.
• Led design reviews and mentored junior engineers, cutting onboarding time by 30% while raising the bar on code quality.
• Partnered with Product to translate complex compliance requirements into phased technical roadmaps.

• Re-architected capacity management to Slot-level granularity in 7 days, shrinking expansion lead time from 4 weeks to 1 week ahead of a high-visibility marketing launch.
• Delivered Real-time Demand Shaping v2 - cut peak-traffic order failures by ~1% (absolute), recovering ~$1k+ GMV per day.
• Led backend engineering for the InsanelyGood pivot, designing high-availability presentation-layer services to support a major product transition.
• Orchestrated end-to-end integration of delivery tracking during a high-stakes migration into the primary Swiggy ecosystem.

Instamart was preparing a major marketing launch and needed to expand into new dark store slots within a week. The existing capacity system managed availability at the zone level - too coarse to allocate individual delivery slots without overcommitting. I re-architected the system to track capacity at the slot level in 7 days flat, cutting expansion lead time from 4 weeks to 1 week and letting the launch proceed on schedule without order failures.

• Re-engineered Trade Allocation workflows to handle 10× the throughput while cutting infrastructure costs by 80%.
• Built a Trade Position Open-Close calculator preventing $10,000+ in projected monthly losses.
• Deployed Kong API Gateway for the Advisor Portal, centralizing auth, rate-limiting, and traffic routing.
• Automated regression testing for Derivatives Exchange Connectivity, cutting release cycles from 8 hours to 2 hours.
• Optimized high-throughput Kafka processors handling real-time trade data streams over the FIX Protocol.