Jan 2022 – Dec 2024 · 2 yrs 11 mos

• Fulfillment 2.0:
• Revamped fulfillment services with new domain models and modern tech stacks.
• Scaled daily processing capacity by 10x, ensuring 99.98% uptime.
• Reduced KTLO from 20% to nearly hands-free operation.
• Launched reusable horizontal platforms, along with achieving the highest-ever seller NPS and lowest promise breaches.
• Payment Service for Sellers:
• Led development of a scalable, high-performance payment service handling millions of weekly transactions.
• Total value of successful transactions crossed 120M.
• Monetized account management, with a future-ready payments solution.
• Organization CoE:
• Led org-wide Dropwizard upgrade, increasing service stability and performance.
• Managed central package repository, enforcing consistency and improving code reuse across teams.
• Spearheaded Istio onboarding to modernize service mesh infrastructure and improve observability.
• Pioneered the “one-click” ephemeral environment project, reducing testing time by 70%.
• Seller Ratecard 3.0:
• Overhauled the ratecard system, reducing rate upload time from 2 days to 4 hours.
• Designed a new, efficient ratecard structure optimized for constant-time evaluation (O(1)).
• Implemented automated access control and approval workflows.
• Enricher Service Revamp:
• Achieved 90% latency reduction and 10x throughput increase via DAG-based processing.
• Optimized resource allocation, reducing infra costs by 80%.
• Tech Solutions:
• Solutioned and reviewed 70+ product requirements spanning 8 teams, delivering well-documented, scalable solutions.
• Ensured timely execution with negligible open issues, resulting in successful project delivery with high Say/Do.
• Training & Mentorship:
• Conducted 4+ training sessions for 300+ devs and leaders on tech, finance, and DS topics.
• Achieved a 4.9/5.0 rating, with later attendees implementing advanced projects that drove business outcomes.
• Other Projects:
• Returns 2.0: Revamped return workflows, boosting efficiency and cutting processing time by 82%.

Jan 2020 – Dec 2021 · 1 yr 11 mos

• Experiences:
• Developed a next-gen Report Center for sellers, enabling near-real-time report generation with Spark, reducing costs by over 95%. Migrated from Vertica to a Spark-based solution on HDFS with 20TB+ data, cutting generation time from 2+ hrs to secs. Implemented a self-serve platform for the business team to onboard reports, reducing onboarding time from 1 week to 4 hours.
• Optimized the Seller Payments page, reducing latency from 40 seconds to less than 1 second. Achieved a 95%+ reduction in CPU and resource usage by implementing ES routing and tuning configurations. Decreased network calls by 80% through a bucket query strategy. Fine-tuned caching patterns and service access patterns to enhance overall performance and efficiency.
• Led the first ES upgrade in the org, to ES v7.x with a 20TB data cluster. Successfully migrated 20+ MySQL clusters to v5.7, handling TBs of data. Kafka -> Pulsar migration, MySQL -> ES migration, pioneered VM -> Kubernetes migration of 11+ services and provided org-wide knowledge transfer with detailed SOPs.
• Developed a T-1 day-based solution for payment orchestration, reducing payment failures by over 99% and enabling the payout process to be triggered a day earlier. This enhancement eliminated contract breaches and provided earlier visibility of payment amounts, significantly reducing seller anxiety. The solution efficiently distributed a load of payments to lakhs of sellers from a 2-hour peak to a non-peak 12+ hour window, improving system reliability and seller satisfaction.
• Published research paper on Default Seller Detection using ML models. Leveraged seller & listing level metrics, and payment trends to predict the likelihood of seller default with an accuracy rate of 92%.
• Proposed and developed a scalable transfer pricing model within 24 hours during a hackathon. The model calculates listing prices based on the final price provided by sellers.
• Other projects: Automation Suite, OSS Vulnerability, Partner Network 2.0

Jul 2018 – Dec 2019 · 1 yr 5 mos

• Experiences:
• Read Scaling of Flipkart RateCard to Sellers: Implemented ElasticSearch Read Layer on top of MySQL to scale to 10x QPS along with bringing down latency to under 200ms. There was a minimum of 6 hops to the DB for a single request with multiple pessimistic locks and transaction.
• ElasticSearch solves for fetching the appropriate agreements for a particular fee calculation based on various parameters like weight, shipping zone, etc. A seller is charged multiple numbers of fees (like platform fee, shipping fee, etc) against each item of an order (to give an idea of scale). Guava caching was used to store Serializable compiled MVEL statements present within the agreements, which were evaluated against the parameters parallelly. This solved the QPS benchmarking targets for 2-3 more Big Billion Days.
• Seller Report Center: Resolved a critical pool connection exhaustion issue, reducing error frequency by 80%. Optimized slow queries, cutting execution time from over 2 hours to under 500 ms, and achieved real-time search performance. Identified and proposed solutions for memory leakage in Hibernate, leading to a 20% improvement in system stability during peak loads.
• Resolved major issues in the report generation process, reducing stuck reports by over 95% from 2,000+ daily. Optimized the system by aligning subreport creation and report generation, improving average processing time to 2-3 hours during high load. Identified and documented failure points, including infinite loops and single points of failure, for faster resolution and enhanced system reliability.
• Data-center Migration: Spearheaded the successful migration of AGM service to Hyderabad with minimal downtime, making it the first service in the organization to migrate without systemic issues. Managed cross-DC latency challenges, reducing latency spikes and ensuring system stability.

May 2017 – Jul 2017 · 2 mos · Chennai Area, India

• Worked with Enterprise Security Solutions team to learn about various Cyber Attacks and the defense mechanisms used by various companies developed by Wipro.