Skip to content

Projects

What I have built.

Two projects of my own, built by me alone and operated by me, plus one cloud project on top of someone else codebase. The detail section opens up if you want to know more.

Masterschool graduation projectLive

Florabase

Smart grow and plant care platform

Florabase supports plant cultivation from balcony growing to professional operations. Users monitor sensor readings such as pH, EC and medium moisture, get alerts when thresholds are breached, document care and generate QR labels. On top of that: AI-assisted diagnosis of diseases and deficiencies from a photo, plus automatically generated cultivation plans.

V2 live, V3 in development, mobile app being built with Capacitor

For companies: details and stack

Tech stack

Frontend (app)
Reactshadcn/ui
Mobile
Capacitor (iOS/Android)
Frontpage and content
AstroDirectus
Backend
PythonFastAPIPostgreSQLREST API
AI
Claude (AI Advisor)Kitwise API (AI Doctor)
Infrastructure and operations
KVM/Ubuntu-ServerDockerCloudflare TunnelCloudflare WAFMinIOCrispGlitchTip

What makes it interesting for companies

  • Decoupled architecture: the React app, a separate Astro content site and the FastAPI REST API can be deployed and scaled independently.
  • Three user tiers covered: beginners, advanced growers, professionals and organisations.
  • AI as the differentiator: photo diagnosis of diseases and deficiencies plus automatically generated cultivation plans.
  • Application and data run on my own containerised infrastructure. Full data sovereignty over the core of the platform.
  • Built for operations: versioned database migrations, tests, error monitoring, with Cloudflare Tunnel and WAF as the protective layer in front.
  • One codebase for web and mobile: the React app ships as an iOS and Android app through Capacitor.
  • On the roadmap: automation and IoT integration so sensors and controllers talk to the platform directly, later for organisations and external systems.
Open source, AGPL-3.0In development

Qurl

Open-source QR code service with dynamic codes

Qurl is a self-hosted QR code service: design codes with colours and a logo, export them as PNG or SVG and point them anywhere. The codes are dynamic and resolve through your own server, so target, expiry and permissions stay changeable. It also includes a vCard generator for digital business cards and branded landing pages with action buttons.

In active MVP development, website coming at qurl.tom-schorn.de

For companies: details and stack

Tech stack

Backend
PHPSymfony 7PostgreSQLDoctrine ORM
Frontend
TwigSymfony UX (Turbo, Stimulus, Live Components)Bootstrap
Core features
endroid/qr-codeSymfony Security (Rollen)i18n (DE/EN)
Infrastructure
DockerPortainer (Git-Stack)self-hosted

What makes it interesting for companies

  • Dynamic codes: target, expiry and permissions stay changeable because the codes resolve through your own server.
  • Fully self-hosted and open source under AGPL-3.0. Full data sovereignty, no vendor lock-in.
  • More than QR: a vCard generator and branded landing pages with action buttons.
  • Accounts, roles and a premium tier for special permissions are part of the design from the start.
  • Deployed as a git-linked Portainer stack, which makes it reproducible and easy to operate.
Cloud Engineering final projectWork in progress

GroceryMate on AWS

Cloud deployment of an existing e-commerce application

An existing e-commerce application built from a Python/Flask backend, a React frontend and PostgreSQL is moved from a local setup into a production-ready AWS environment. It is not about new features but about the questions behind them: how to build it reproducibly, run it securely, monitor it and keep the cost sane.

My part

The GroceryMate application is not mine. It was built by Alejandro Roman Ibanez as part of the Masterschool programme. I work on a fork and handle only the cloud and operations side: containerisation, infrastructure and deployment on AWS. The application code is the starting point, not my contribution.

Containerisation started, AWS architecture in progress

For companies: details and stack

Tech stack

Application (given)
PythonFlaskReactPostgreSQL
My part
DockerAWS

What makes it interesting for companies

  • The realistic case in day-to-day work: not building greenfield, but understanding someone else code and making it production-ready.
  • A clear separation between application and infrastructure. The template stays untouched, the operational layer wraps cleanly around it.
Lab experiment

Quantum Field Formation

Photography without a camera, using only the residual field of your screen

Every screen emits scattered photons that reflect off everything in the room, including you. Within a coherence window of a few nanoseconds, an image can be reconstructed from them. No camera required. Just press the button.

No camera is requested. Promised.

More on GitHub

Further contributions and code samples.

github.com/tom-schorn →