I build Android apps and the distributed backends they depend on — currently shipping autonomous drone software at Complago, with my attention increasingly on artificial intelligence and serverless cloud architecture.
Software Engineer
Complago
I'm a software engineer with a Computer Science degree from the University of Crete and production experience in Android development and distributed drone systems. At Complago I work on an autonomous flight application in Kotlin and Jetpack Compose — mission control, telemetry, live video, and cloud-connected workflows on real DJI hardware.
My thesis pulled at the other end of the same problem: what happens when a drone drops out mid-mission. I modelled the mission as portable execution state that can be checkpointed and transferred between drones, built the FastAPI backend and the Python/MAVSDK agents that coordinate it, and measured how long the gap actually lasts.
I also spent two years teaching — tutorials, lab hours, and grading for Calculus I & II and Object-Oriented Programming in Java. Explaining a concept to a room that hasn't seen it before is the fastest way to find out whether you actually understand it.
Reading Russell & Norvig's Modern Approach, 4th edition.
Event-driven services on Azure and AWS.
Consistency and recovery under partial failure.
A short view of what I've finished, what I'm building now, and what I'm deliberately learning next.
Four years across systems, algorithms, databases, and software engineering, alongside two years assisting on Calculus I & II and Object-Oriented Programming in Java.
Fault-tolerant handover between drones: lease-based ownership, checkpoint/resume, and crash recovery, evaluated across four scenarios. Supervised by Prof. K. Magoutis.
Building the drone pipeline behind Scylax Control — waypoint missions, telemetry, live video, and cloud upload on DJI hardware, in a production codebase.
Working through Russell & Norvig's Artificial Intelligence: A Modern Approach (4th edition), from search and probabilistic reasoning to machine learning — and looking for places it fits the autonomy problems I already work on.
Moving my backend work toward managed, event-driven services: functions, queues, and managed data stores, and the trade-offs in cost, cold starts, and operational surface against the containers I run today.
Production Android app for autonomous drone missions.
The DJI-based flight pipeline at Complago: waypoint mission execution, telemetry handling, real-time video streaming, and cloud photo upload. Built in Kotlin and Jetpack Compose, shipped through the full cycle from feature work to debugging and delivery.
Keeping a drone survey alive when the drone flying it goes down.
Thesis work: a surveillance mission modelled as portable execution state that can be checkpointed and handed to a second drone without data loss. Three layers — PX4 SITL/Gazebo for flight, Python/MAVSDK agents for execution, FastAPI + Docker for shared state — with lease-based ownership and crash recovery. Measured across four scenarios: 39s for a graceful handover, 81s for crash recovery, with sensor deviation held to ±0.25°C.
Offline-first inspection platform that survives process death.
Field technicians fill in forms and attach 5–40 MB photos with no connectivity. Resumable chunked uploads commit progress to Postgres before acknowledging, so an app killed mid-transfer resumes exactly where it stopped. Delta sync with optimistic locking and device-minted idempotency keys make every retry safe.
Python agents that fly PX4 drones through segmented surveys.
The execution layer of the thesis: MAVSDK agents that claim mission segments, emit heartbeats, checkpoint their progress, and take over automatically when the active drone stops responding.
Two data-driven web apps built on Java Servlets and MySQL.
University coursework (CS359, CS360): an event management system and a fire department dispatch system, each with a relational schema, RESTful endpoints, and a server-rendered front end.
More on GitHub.