CaidenClinic
Shipped a production clinic-ready system in 4 weeks with a 3-person AI-native team and no dedicated PM.
Client / Project
CaidenClinic is an internal proof-of-work product for Meta.Dev — a full patient intake, scheduling, and billing system built to validate that a small AI-native team can replace a traditional 6-person product squad for a clinic-grade SaaS.
Problem
The reference customer (a small primary-care clinic) had outgrown spreadsheets but couldn't justify the $80k+ quotes from vendors. Off-the-shelf EMRs were over-featured, expensive, and slow to onboard. They needed something focused: digital intake forms, appointment scheduling, automated reminders, and Stripe-backed billing — built in weeks, not quarters.
Approach
Three Meta.Dev engineers, no separate PM or designer. We used agentic workflows for the boilerplate (CRUD endpoints, form schemas, migration scripts), kept human attention on the parts where judgment matters (clinical workflow modeling, HIPAA-adjacent data handling, payment flow), and shipped behind a feature flag every day.
Architecture stayed small: Next.js for the patient-facing portal and admin app, FastAPI for the clinical workflow service, Postgres for state, Stripe for billing. No microservices, no message bus — one app, one database, fast feedback loops.
Outcome
The clinic was running production intake within four weeks. Two months in, it's processing a few thousand appointments a month with no engineering touch — which is exactly the point. The product proves the thesis: a small AI-native team, embedded directly into the problem, can match traditional squads at a fraction of the cost and timeline.