At TSS Consultancy Pvt Ltd, I designed and scaled a Java Spring Boot KYC verification microservice processing 50,000+ daily API requests, with an average response time of 1.5 seconds.
I also built a document-matching and quality-check engine that added 40,000+ daily requests to the platform. It uses fuzzy-matching confidence scores across name, address, and DOB attributes, with client-specific response filtering.
I improved backend inference performance by converting PyTorch and PaddleOCR models to ONNX, reducing CPU inference latency by 25%. On a Flask service, I addressed recurring memory and CPU issues with Gunicorn multi-worker configuration and threshold-based restarts.
I implemented tenant configuration and JWT-based tenant claims, and built a Spring Boot rate-limiting framework using AOP. I also developed an internal code-review service integrating the OpenAI API with Azure Repos, reducing manual review turnaround time by 40%.
