Michael Biswas

Led development of an AI-powered patient communication platform for automated appointment scheduling, reminders, and inquiries using conversational voice workflows. Improved onboarding and patient interaction flows, increasing platform conversion from 55% to 69.3%, and migrated backend APIs from REST to GraphQL to support faster AI-driven workflows.

Built high-throughput AI inference pipelines with FastAPI, Python, and TensorFlow, reducing model response latency by 38% through async execution, caching, and GPU utilization. Developed model lifecycle tooling and optimized production deployment by reducing release time from hours to under 15 minutes via CI/CD automation and Kubernetes-based orchestration.

Developed scalable financial analytics modules with Django and React, automating balance sheet, P&L, and cash flow reporting for 50,000+ users. Built PyTorch-based expense categorization improving classification accuracy by 27% and reduced Azure infrastructure costs by 18% while scaling microservices and dashboards.

Contributed to core React and React Native frameworks by implementing performance improvements and critical bug fixes to enhance rendering efficiency and developer experience. Improved React Native’s JavaScript-to-native bridge for better communication speed, memory handling, and stability, and expanded test coverage and tooling for smoother feature adoption.

Earned a Bachelor's degree in Computer Science at the University of California, Berkeley from 2009 to 2013.