Team build · full-stack IoT
- —Office air quality, from sensor firmware to a role-based dashboard
- —Four-layer stack: C/C++ → Python edge → Java → React, front and back in a 6-person team
- —Privacy rule enforced twice: edge + server validation
Office air quality, sensor to dashboard
Sensor stations in the rooms, an edge box that filters and forwards, a Spring backend and a role-based React dashboard. Six people built it; front and back end are mine.
The system
Four layers, each a different stack, all talking to each other:
- Arduino Nano 33 BLE (C/C++):Sensor stations reading temperature, humidity and air quality from a BME688, advertising over Bluetooth Low Energy.
- Raspberry Pi edge (Python, bleak, SQLite):Reads each station every 10s, applies a 5-sample sliding median to filter sensor noise, and enforces a privacy rule locally before forwarding.
- Spring Boot (Java 25):Central server — user management, role-based visualizations, climate scores, threshold warnings.
- React 19 / TypeScript:Dashboard, with a TypeScript API generated from the backend's OpenAPI spec.
Privacy by design
The spec carried a long list of privacy rules and exceptions. Each one is enforced twice — once on the edge box before anything leaves the room, once again on the server — so a single misconfigured layer cannot leak a reading.
C/C++BLEPythonSpring Boot Java 25React 19DockerPostgreSQL
More work
Event Management PWAEvent-management PWA, built and shipped solo.RIS MCP ServerMCP server that opens Austrian federal law to LLMs.Industrial Printhead-Cleaning SoftwareControl software for an industrial printhead-cleaning rig.What's possibleWhat comes out when an idea gets built all the way, with nothing in the way.
AI automation