At WAMO LABS, I architected a Python-based AI/RAG pipeline that uses document chunking, vector retrieval, and multi-agent evaluation to extract and validate information across 100+ business fields. I also implemented an LLM-as-a-Judge approach with specialized agent personas and majority voting.
That work reduced document review and processing time from approximately 12 hours to under 1 hour. I also designed AWS SQS-based asynchronous communication between three microservices, with dead letter queues to help isolate failures and support reliable communication.
At Techfy, I restructured a Django application for multi-tenant organizations and implemented role-based access control. I developed an administrative application with Next.js and deployed it on AWS EC2, while performance optimizations reduced page load times by up to 40%.
