• Led the successful migration and deprecation of legacy APIs, enabling 100% bug resolution rate and reducing development time for fixes/features by 50%.

• Scoped and guided an intern project to integrate a caching layer between service and database, providing continuous technical mentorship.

• Redesigned Elasticsearch sharding and indexing strategy, improving reliability and performance of the device metadata read APIs. Notably, achieved a 20% reduction in P(50|90|99) latency and a 95% decrease in customer and API service-related issues.

• Designed and developed a high-performance API that enables seamless retrieval of metadata for newly registered Alexa Devices, catering to the needs of the hospitality industry (hotels, offices, hospitals) with sub-100ms response time.

• Re-architected and created an Elasticsearch (ES) cluster leveraging AWS CDK to establish a robust access mechanism to the ES cluster via a bastion host, facilitating streamlined debugging, testing, and efficient data management through a command-line interface.

• Orchestrated and executed comprehensive operational preparations for Q4, strategically scaling the hosting environments in anticipation of peak traffic of 180,000 transactions per second (tps).

• Co-developed a scalable backend architecture in a 2-developer team of 4, delivering core infrastructure and a GraphQL API layer in under two months. The system handled complex business logic and enabled efficient, client-specific data access, serving as the foundation for future feature development.

• Implemented a role-based authentication and authorization system to secure application endpoints and ensure proper access control across user types.

Stack: Node.js, Express, GraphQL, MongoDB, Docker, Kubernetes

• Generated 100,000s of synthetic images for object detection model training, improving the model’s accuracy through targeted augmentation and dataset design.

• Optimized the preprocessing pipeline, reducing data prep time from over 10 minutes to under 2 minutes by streamlining image transformation and I/O operations.

Stack: Python, TensorFlow, OpenCV, Shell Scripts