Staff Software Engineer
Capsule
April 2025
-
Present
- Designed and implemented an LLM-powered AI agent system using Pydantic AI that accelerated prior authorization processing by 70%, directly improving patient access to medication.
- Designed event-driven task routing system using microservices to distribute prescription processing failures to pharmacists and technicians based on real-time availability
- Improved database performance leading to 21% reduction in CPU usage, 35% reduction in IOPS and 50% reduction in costs.
- Backend - Python, pytest, Django, PostgresSQL
- Frontend - Typescript, Next.js, React, Vite
- Infrastructure/Other - Docker, Kubernetes, AWS (Lambda, ElastiCache, EKS, RDS, SQS/SNS), Shell, CI/CD pipelines, Sentry, Grafana (Prometheus, Loki)