NextFix AI — Field-Service SaaS Platform

Overview

NextFix AI is my deepest personal R&D project — an end-to-end field-service platform (think dispatch, technicians, jobs, invoicing) built the way I would architect a real product: a fleet of focused microservices, typed contracts shared across every surface, and native mobile apps for both customers and technicians.

Architecture

The backend is 14 independently deployable NestJS services behind an API gateway — auth, users, organizations, technicians, dispatch, service-requests, diagnostics, files, invoices, analytics, and notifications. Services communicate over a message broker, share a Zod-validated contract layer, and each own their slice of PostgreSQL. A later iteration (v2) moved coordination to NATS and added Cedar policy-based authorization, JWKS token verification, and Prometheus metrics.

Highlights

• AI diagnostics service integrating the Anthropic SDK to triage incoming service requests.
• Queue-driven workflows with BullMQ for notifications and long-running jobs.
• Two Expo (React Native) apps — customer and technician — plus a Next.js admin dashboard, all sharing the typed API client.
• Supabase auth as the identity provider, JIT-mirrored into each service.

Stack

TypeScript · NestJS · PostgreSQL · Redis · NATS · BullMQ · Supabase · Expo / React Native · Next.js · Docker