Architected and led the phased, strangler-fig migration of a legacy PHP EMR platform onto a Next.js/NestJS microservice stack — designing the infrastructure and technical architecture from scratch, and leading a small team through it. RabbitMQ handles inter-service communication; BullMQ and Redis drive the cron-based activity-logging pipeline, backed by MongoDB. Alongside the migration, I led the team's shift to an AI-native development pipeline — Cursor and a custom Claude Code skill system (requirement-interview → clean → verify → commit end to end).

• A project scoped for ~2 years shipped in significantly less time
• Strangler-fig phased cutover with zero migration downtime
• AI-native dev pipeline adopted end-to-end across the team

Designed a lead-generation platform that combines automated NPI registry data enrichment with an AI layer, feeding directly into a sales pipeline built for the lead-gen and SDR teams — from raw registry data to a qualified, sales-ready lead with no manual handoff.

• Automated NPI registry enrichment that previously required manual lookup
• Sales pipeline integrated by default for automated SDR workflows
• Built and owned by the lead-gen team without engineering hand-holding

Full architecture and development of a remote EMR integration synchronizer — a WebSocket-based remote agent that keeps client-server EMR data in sync and exposes remote actions (start, stop, update, schedule) back to the integration itself.

• Real-time sync between client servers and central EMR integration
• Remote-controllable agent (start/stop/update/schedule) from a central console
• Also connects the team's AI development pipeline to Zoho and Jira

An intelligent shipyard-management web and mobile application built to improve operational efficiency, developed working directly with the client to gather requirements and deliver a working prototype.

• Delivered a client-validated prototype from direct requirements gathering
• Cross-platform web and mobile from one codebase