Senior Full-Stack Python Engineer (FastAPI + React) Take Over a Ready IoT Facilities Platform

We are a Saudi total facilities-management (TFM) company. We have built our own CAFM system - and our operations will run on it daily: our supervisors dispatch work from it, our technicians close jobs on their phones through it, and our clients log in to it to track their sites. We're hiring one senior engineer to take it to production and own it long-term. This is first and foremost an operations platform, not an IoT project. The system of record for our maintenance business is the CAFM core: - Work order lifecycle — client requests, approval, assignment, execution with enforced completion discipline (notes, time, failure codes, photos), supervisor close-out - SLA engine — four contract gates (acknowledge → attend → contain → resolve) per priority, hold-aware clocks, and an evidence export we use to settle disputes with clients - PPM planner — 52-week SFG20-style schedule, statutory/mandatory compliance classes, auditor-grade compliance math - Technician mobile app — installable PWA, offline-capable with honest timestamps, guided execution journey with H&S and quality gates - Dispatch cockpit — breach-prevention board, unassigned pool, live technician load - Client portal — role-separated views, request/approval loop, monthly bilingual (English/Arabic) PDF reports with contract KPIs - Supporting modules — asset register with QR codes, audits & condition surveys, permits to work (LOTO/hot works), HSE incident register, snagging/handover punch-lists, stores & job costing On top of that core sits our differentiator: LoRaWAN sensors (Milesight) stream into a rules engine that raises alerts and work orders automatically, plus a predictive layer that projects failures before they happen. You'll maintain this too — but the daily heartbeat of the business is the CAFM. Practical implication for you: uptime, data integrity, and careful change management matter more than new features. When our operation depends on the work-order board at 7am, "it broke after my deploy" is not acceptable. We expect staging discipline, backups with tested restores, and small, reviewed changes. The stack (already built, feature-complete, 106 automated tests): - Backend: Python 3.12, FastAPI, async SQLAlchemy, PostgreSQL 15 + TimescaleDB, Redis + arq workers, Alembic - Frontend: React + TypeScript + Vite + Tailwind, PWA, fully bilingual EN/AR (RTL) - Infrastructure: Docker Compose, documented production-readiness checklist and runbook Your first engagement — the production gate (fixed milestone): - KSA-region hosting, TLS, secrets management, monitoring, log retention - Database backups with a proven restore drill - CI with an isolated test database - Wire production sensor webhooks, SMTP, and messaging provider Then ongoing (monthly retainer): - Operate the platform our company runs on — uptime, monitoring, careful releases - Onboard new client sites (assets, PPM plans, users, SLA matrices) - Incremental features from an existing documented roadmap Must-haves: - 5+ years full-stack, with systems you personally deployed and operated in production - Strong FastAPI + async SQLAlchemy + PostgreSQL - Docker-to-production: TLS, secrets, backups, monitoring - Comfortable in React + TypeScript; Redis + background workers Strong pluses: - CAFM / CMMS / field-service / ERP domain experience — you know what a work order, PPM schedule, and SLA mean operationally - TimescaleDB or time-series experience; IoT/LoRaWAN/webhook ingestion - Arabic or RTL product experience; GCC hosting and PDPL awareness How we'll start: Shortlisted candidates get a paid trial milestone: clone the repo, stand the stack up locally (one command), run the test suite, and deliver a 1–2 page review of our production-readiness checklist with your deployment plan, timeline, and cost. Real paid work — it's how we pick the finalist. Terms: - NDA and full IP assignment; code stays in our repository, infrastructure in our accounts - One accountable engineer, not an agency team; long-term preferred - Minimum 3 hours/day overlap with Saudi Arabia (GMT+3) Screening questions (answer all — applications without them are declined): 1. Describe a business-critical system you operated in production. What failed unexpectedly and what did you do? 2. Have you built or maintained a CMMS/CAFM, field-service, ticketing, or ERP-type system? Describe the workflow it managed. 3. How would you back up PostgreSQL/TimescaleDB and prove the restore works? 4. What is your step-by-step approach to taking over a codebase you didn't write? 5. Start your application with the word PULSE.