Oct 2024 – Jan 2026 · 1 yr 3 mos · Bengaluru, Karnataka, India · Hybrid

• Own and maintain services handling 500K+ requests per minute with average latencies of 100ms and 0 downtime, powering major monetization pipelines across the platform. Designed for horizontal scalability, graceful failovers, and observability, with deep focus on performance tuning and GC optimizations.
• Designed and implemented a high-performance microservice for Zepto’s ad product catalog(Ads Catalogue Service) , enabling real-time stock and product updates with 30ms P99 latency and eliminating reliance on core service APIs.
• Designed and launched a config-driven ad-ranking (Ads Ranking Service) microservice supporting pagewise and weight-based scoring with multi-attribute relevance experiments, removing the need for developer intervention.
• Integrated advanced monitoring and alerting pipelines (Grafana, Prometheus etc.) to ensure 99.99% uptime and quick mitigation of anomalies. Proactively tracked system KPIs, identified bottlenecks, and improved overall system reliability and developer experience.
• Responsible for design, development, deployment, and monitoring of core ad-serving components.Regularly perform code reviews with emphasis on test coverage, design clarity, and scalability. Set a high bar for code quality and best practices across the team.

Apr 2024 – Sep 2024 · 5 mos · Bengaluru, Karnataka, India · Hybrid

• Implemented a high-performance personalized ads pipeline for the home page vertical feed, integrating real-time user context, ranking models, and ad eligibility logic. Designed scalable APIs and caching layers to support low-latency delivery at scale. Resulted in a 22% increase in CTR (0.50% vs. 0.41%) and 3x growth in revenue per impression.
• Designed a modular backend for PLA ads, exposing RESTful APIs for campaign configuration, inventory ingestion, candidate generation, and real-time auctions. Built a highly concurrent ranking engine with multi-stage filtering and relevance-based scoring. Sustains majority of ads traffic with <70ms response times.
• Architected and led the development of a dynamic runtime bidding engine(RTB) with multi-strategy support and intelligent lexical score balancing, driving a 13% increase in Zepto’s ad revenue.

Jun 2023 – Apr 2024 · 10 mos · Hybrid

• Part of SC and MOJ Ads Relevance team where work mostly revolves around auction , bidding, ranking and relevance of Ads .
• Worked on the promoted posts architecture and end to end delivery of the project which increased the swvp to ~15% from ~5% and also HTC posts which we have seen 50% uplift in engagement along with cost gains .
• Led the onboarding of ARS Hamsa architecture which is in-house in-house vector search platform for ML teams/services .
• execution of Dependency Injection modifications in the ARS service, ensuring a flawless implementation which increases the unit test coverage of the service by 40%.
• Worked on Data Observability in the system and built Feature Health.

Jun 2022 – Jun 2023 · 1 yr · Hybrid

• Worked on Data Observability in the system and built Feature Health.
• Led the onboarding of TARDIS (in house feature store) for ads use case .