At Kidture Health Inc., I built a LangGraph-based multi-agent clinical AI system that routes voice and text inputs to specialized medical agents, with safety guardrails for potentially acute symptoms. It reduced parental logging friction by 60%.
I designed a FHIR R4 clinical data architecture that maps caregiver-reported symptoms, medications, and feeding inputs to standardized terminologies. I also built a speech-to-text pipeline and data synthesis workflows that combine chat data with Apple HealthKit telemetry.
At Freshworks, I worked across AI platform orchestration, backend systems, and infrastructure. I architected a RAG ingestion pipeline for Freddy AI and improved incident response by 40% through monitoring, PostgreSQL query optimization, and automated alerting.
Earlier at Freshworks, I developed event-driven processing systems, database scaling and migration approaches, and React dashboards. My projects include Rust middleware for LLM safety and a healthcare assistant that won the Rutgers Health Hackathon.
