Manaswini Upadhyayula

At AI Vectors.org, I architected a three-account AWS environment for a HIPAA-compliant healthcare AI platform designed to scale to 10,000 concurrent users. I restored platform availability during five concurrent production outages by resolving NGINX misconfigurations and redesigning least-privilege IAM access across 16 AWS services.

I re-architected the application server from Uvicorn to Gunicorn, resolving the timeout conflict behind 89.6% of production 502 errors. I also engineered Redis and DynamoDB DAX caching that reduced peak database queries from 10,200 to about 100 while preserving PHI handling.

Through ElitTech, I deployed a fully serverless AWS platform with Terraform, GitHub Actions, CloudWatch, and SNS alerting, reducing deployment time from 15 minutes to under 30 seconds. I enjoy owning infrastructure from IaC and CI/CD through monitoring, security, and operational debugging.

I also build full-stack products independently, including an AI Gym Trainer using TensorFlow.js and a microservices e-commerce platform running Node.js, Java, Python, RabbitMQ, NGINX, MongoDB, MySQL, and Redis.