Efstratios “Stratis” Avagianos

A Wi-Fi connected emergency button designed for elderly users who cannot reliably use smartphones. With a single physical press, the device sends instant alarm notifications to predefined contacts.

The system is built around an ESP32 microcontroller and includes two physical buttons, a status LED, and a battery-powered circuit. Communication is handled over Wi-Fi, allowing alerts to be triggered immediately without any complex user interaction.

A companion Android application, developed in Android Studio, manages user accounts, authentication, and real-time communication using Firebase. Firebase is used both for device-to-app messaging and as the primary database.

The enclosure was custom-designed by me in Fusion 360 and manufactured using a 3D printer. A planned improvement is the integration of a fall-detection sensor to automatically trigger alerts.

A personal self-hosted environment running multiple websites and services on a Raspberry Pi. The setup uses Docker containers for service isolation and NGINX as both a web server and reverse proxy.

This project provided hands-on experience with real deployment constraints, including networking, port management, service separation, and long-term reliability. All configuration and maintenance is handled manually without managed hosting services.

A gamification layer built in Unity that reacts dynamically to real-time data streams. The application consumes live JSON data and updates visuals and behavior based on incoming values.

Instead of static dashboards, this project explores how live information can be communicated through motion, feedback, and interactive systems.

A remote automation system that allows my PC to be powered on from anywhere in the world. A Telegram bot command (/wake) is sent to a service running on my Raspberry Pi, which then triggers the PC startup sequence.

This project combines messaging APIs, networking, and automation logic to solve a practical everyday problem: remote access without physical presence.

A group university project (team of three) focused on protecting children from harmful online communication. The app monitors incoming Telegram messages and detects aggressive or dangerous content.

When such content is detected, the system immediately alerts the parent through a companion application and forwards the message for review. The system uses AI-based text analysis for real-time classification.

Backend services are containerized using Docker, while the front-end is built with Flutter. The project emphasized collaboration, security awareness, and real-time processing.