May 2022 – Present · 4 yrs

• API optimizations reduced the impact of certain APIs on server utilization from 30-40% to less than 5%, improving overall performance of Service (Elastic APM).
• Developed highly responsive, secure and scalable asynchronous services using Spring Webflux, Spring security, Reactive Mongo, and Java (65% throughput improvement).
• Optimized mongo queries, implemented caching, and minimized expensive I/O operations and database queries, reducing monthly costs by 34%.
• Automated and optimized file processing through serverless cloud features like AWS Lambda using Serverless framework, improving workflow efficiency by 40% and reducing manual labor by almost 95%.
• Identified performance bottlenecks and optimized system configuration through load testing and contract testing, enabling the application to support up to 1000 concurrent users.
• Developed a denoiser application leveraging an ML model to clean up images. Integrated RabbitMQ and a Node app for sequential image processing, increasing order conversions by 8% and cost reduction on AWS EC2.
• Developed an in-house MOM system for client meetings, leveraging a local LLM for summarization and templated output. Automated email triggers to relevant stakeholders, resulting in an approximate 2% increase in order conversion.
• Leveraged concepts from "Automatic Interior Design in Augmented Reality Based on Hierarchical Tree of Procedural Rules" to design a rule-based, sequentially procedural system for auto-populating objects within a given space. Focused on system architecture, design principles, and rule hierarchies, advancing this as a proof of concept.