Experience

US healthcare SaaS platform serving 50+ clinics, 5,000+ patients, ~300 active staff users, and ~500 studies/day. Team of 7 engineers. End-to-end architecture ownership across patient intake, authorization, RIS/PACS, billing, payments, and AI-enabled operations.

• Owned architecture and delivery of 5 core platform modules (RIS, patient intake, e-forms, billing, payments) — the operational backbone for all clinical administrative workflows across 50+ US clinics
• Designed payment-safe billing architecture with idempotent Worldpay integration, webhook reconciliation, and transaction locking
• Built event-driven integration layer (AWS Lambda, SQS, DynamoDB, SNS, Websocket, SignalWire) with retry classification, dead-letter queues, and correlation IDs — reduced missed notifications to near-zero for time-sensitive clinical alerts
• Owned RIS/PACS integration architecture, designing workflow orchestration across radiology imaging, scheduling, and reporting systems serving ~500 studies/day
• Designed and shipped AI-assisted ops platform (async LLM pipeline with human-in-the-loop approval via SQS, Lambda, S3, Websocket) — reduced median case resolution time; the same ops team now handles more volume without a headcount increase
• Integrated AI-driven job classification into booking workflows with deterministic fallbacks and model-tier routing, reducing LLM costs while maintaining suggestion quality for regulated environments
• Reduced platform load times by 30% through frontend bundle optimization and backend query tuning, directly improving clinician workflow efficiency during peak hours
• Introduced testing standards and code review practices across the team

AI-driven B2B SaaS platform handling loan processing, vendor onboarding, and document classification & validation workflows for enterprise clients.

• Engineered backend services for an AI-driven B2B SaaS platform handling loan processing, vendor onboarding, and document classification & validation workflows
• Replaced hardcoded workflow logic with declarative orchestration (Camunda + Node-RED), eliminating code deployments for workflow changes and reducing new client onboarding configuration from months to weeks
• Built async document processing pipeline (PDF conversion, OCR, structured extraction) that automated manual validation steps and improved document onboarding throughput
• Optimized data retrieval using advanced MongoDB aggregation pipelines, reducing API response times for data-intensive reporting endpoints
• Led schema design and dynamic workflow configuration using Camunda, enabling non-engineering teams to configure business processes

Mentored 30+ junior developers in full-stack web development (React, Node.js, MongoDB, DSA) and system design fundamentals.

• Conducted daily code reviews, led architecture workshops, and provided structured feedback on capstone projects
• Multiple mentees secured engineering roles at product companies within 3 months of program completion