● Led architecture and delivery of a high-concurrency live commerce platform, enabling a seamless shopping experience that scaled to 100K+ concurrent purchases per session and drove a 3x increase in order volume.

● Standardized version-controlled command execution for safe event replay and reliable recovery across deployments, enabling deterministic state rebuild in rollback scenarios.

● Architected a resilient order execution pipeline using Redis (for inventory preloading and write-through caching) and Amazon SQS (for traffic shaping and eventual consistency), achieving <200ms p99 latency under peak load.

● Designed a fault-tolerant payment layer by integrating Stripe with Spring Boot, handling asynchronous payment lifecycle events (success, refund, failure) with secure webhook validation and idempotent processing.

● Served as primary on-call for order/payment pipelines during peak sales events, driving high-severity triage across teams, implementing automated rollback tooling, and institutionalizing post-mortem best practices to enhance system resiliency org-wide.

● Wrote unit and integration tests for critical order and payment flows using JUnit and Mockito, ensuring 85% test coverage and preventing regressions in high-concurrency scenarios during peak traffic events.

● Replaced RESTful APIs with GraphQL to enable flexible, fine-grained data access for multiple vendors, allowing customized queries and mutations while reducing over-fetching and simplifying integration workflows.

● Led the design of a modular blockchain ingestion system for data teams to collect data from ETH, BSC, and Polygon, with per-chain Spring Boot microservices consuming on-chain data via WebSocket and JSON-RPC polling for fallback and reorg recovery.

● Built a CRON-based multi-level data aggregator with daily to yearly rollups. Applied Memcached for caching, write-behind for daily writes, and double-delete strategy for long-term consistency, improving cache hit rate to over 85% and reducing RDS query load by 40%.

● Developed a resilient observability stack using AWS CloudWatch and DataDog to monitor node health, block lag, ingestion throughput, and P99 API latency. Introduced alert-based auto-recovery workflows, reducing mean time to resolution (MTTR) by over 65%.

● Built a high-throughput transaction search API in Java with JPA and partitioned PostgreSQL, supporting filtering by chain, address, and time. Tuned indexes and queries to reduce latency by 60%..

● Worked with the data team to gather feedback and improve ingestion and debugging tools, adding 10+ usability features and reducing issue resolution time by 40% across ETH, BSC, and Polygon pipelines.

● Architected and implemented the backend infrastructure for a short-video platform using Node.js and MongoDB, integrating Azure services (Blob Storage, AD B2C, CDN, Media Services) for identity, storage, and content delivery. Designed for scale and low-latency access under high concurrency.

● Migrated video ingestion from AWS to Azure Media Services for native transcoding and preview generation. Used dual Azure Blob Storage containers (staging + permanent) with SAS access and cleanup policies, reducing review time by 90%.

● Led CI/CD modernization with GitHub Actions and Azure DevOps Pipelines. Integrated feature flagging (Unleash) for controlled rollouts, blue-green deployments, and environment-based toggling across staging and production.

● Developed and maintained automated regression tests for video upload, preview generation, and CDN playback flows. Collaborated with QA to diagnose production issues and improve end-to-end reliability.

● Developed React components for the short-video platform’s upload, playback, and user interaction flows, integrating with Azure Media Services and Blob Storage APIs, improving UI responsiveness and reducing user-reported upload issues by 60%.

● Wrote unit tests for backend services using Jest and Sinon, validating video upload, transcoding, and access logic, which improved test coverage and prevented regressions during CI/CD rollouts.