NutriFlow · Nutritionist Patient Platform
A full-stack web platform for nutrition professionals to manage patients, run physical assessments, build diet plans, and accelerate clinical workflow with AI-assisted support.
Open live platformOverview
NutriFlow is a modern, responsive web application designed for nutritionists who need structured operations at scale. The platform centralizes patient records, anamnesis, physical assessments, and diet plan workflows in a single interface, reducing manual overhead and improving follow-up quality.
Context
Many nutrition professionals still rely on spreadsheets, fragmented notes, and generic chat tools for clinical workflow. That fragmentation creates rework, weak historical tracking, and inconsistent patient follow-up.
Problem
Build a production-ready web platform that supports the complete nutritionist workflow end-to-end: patient onboarding, assessments, diet planning, progress monitoring, and guided assistance for faster decision-making.
Responsibilities
- ›Designed and implemented the App Router frontend architecture in Next.js with TypeScript and Tailwind CSS
- ›Built typed forms for intake, physical assessments, and diet planning with React Hook Form + Zod validation
- ›Integrated Supabase for authentication, relational persistence, and secure document/media storage
- ›Implemented server-side and client-side data flows with TanStack Query for cache consistency and responsive UX
- ›Developed AI-assisted copilot flows to help nutritionists generate and refine plan drafts with better speed
Technical decisions
Next.js full-stack architecture with App Router
Used a single Next.js codebase to coordinate server-rendered pages, server actions, and API endpoints, reducing integration friction while keeping routing and data boundaries explicit.
Supabase as managed backend foundation
Adopted Supabase (PostgreSQL, Auth, Storage) to accelerate delivery of secure core infrastructure while preserving a clear relational model for patient and assessment history.
Schema-driven forms with runtime validation
Standardized form contracts using React Hook Form + Zod to enforce consistent input validation, improve developer confidence, and prevent invalid clinical records from entering the system.
Architecture
Modular feature-oriented web architecture with typed domain contracts, reusable UI primitives, and a clean separation between presentation, validation, and persistence. CI/CD is automated with GitHub Actions and production delivery runs on Vercel.
Challenges
- ›Modeling a flexible data structure for patient history, measurements, and evolving diet plans
- ›Maintaining fast, reliable form UX in large multi-step clinical workflows
- ›Designing AI-assist interactions that are useful in practice without reducing professional control
Results
Consolidated the nutritionist workflow into a single web interface with clearer patient history traceability
Reduced time-to-create and update diet plans through reusable form and template flows
Established a scalable deployment pipeline with GitHub Actions + Vercel for reliable iterations
Learnings
- ›Clinical workflow products require strict data modeling discipline early to avoid costly migrations later
- ›AI copilot features deliver the best value when designed as guided assistance, not full automation