Working as a Full-Stack Developer,Frontend :- ReactJs, HTML, CSS, Next.jsBackend :- Python , Node.jsDatabase :- MySql , MongoDbLed the creation of a highly scalable architecture (v1) tailored for processing automobile images, utilizing Redis, high-throughput SQS, CloudWatch, Node.js, Python, AWS Lambda, AWS Gateway, and MySQL. Completed within 1.5 weeks, it included a pre-calculation architecture and a custom forecasting algorithm for precise processing time predictions and fair task prioritization. This boosted processing capacity from 50-60 images per minute to over 18,000 images per hour during peak times, far exceeding the target of 3,000-4,000 images per hour in optimal conditions.Enhanced the security of public image processing APIs by implementing invisible reCAPTCHA trust tokens, strategic throttling, and code obfuscation. This led to a dramatic reduction in scam processing from 63,327 to zero instances and verified 276K requests, identifying 1% as suspicious since its inception.Launched the development of a web SDK for image processing, managing over 300,000 images within the first 3 months of release. This comprehensive solution included front-end (React) and back-end (Node, Python) development, supported by Redis, SQL, MongoDB, and PostHog. Real-time monitoring with Kibana, Elastic APM, and Slack ensured top performance and user engagement. The web SDK generated 44.77% of revenue in the food category.

Led the Java coding efforts, encompassing comprehensive testing and deployment of the OPUS application on development servers. Oversaw the verification of bundling and validated the incorporation of my individual contributions.

Analyzed the 'baselineWavelet' package associated with 'R' programming language andhelped in its conversion to 'C++', for faster execution and compilation of thepackage. This package is used to calculate peak values in signals.