About

My Story

I’ve always gravitated toward the intersection of software, systems, and intelligence.

I started in embedded systems and networking, writing C for routers and debugging multicore APIs at Ekinops. Before that, I built autonomous racing electronics and PCB designs for a Formula SAE team in Greece, working on everything from power electronics to microcontrollers.

Over time, my curiosity moved upward in the stack: from the electrons on a PCB, to the firmware, to the operating system, to the data pipelines feeding ML models, and finally to the behaviour of AI systems themselves.

This path eventually led me to applied LLM work — first through my master’s at KU Leuven, then through a collaboration with GSK on multi-agent systems in Pharma, and now through designing high-throughput data and ML frameworks for autonomous vehicles at Wayve.

I didn’t “switch fields”; I just kept climbing the abstraction ladder.

What Drives Me

I like hard, ambiguous problems that sit between disciplines — the kind of challenges where no single stack (embedded, systems, data engineering, or ML) is sufficient on its own.

I’m motivated by:

• Building frameworks and tools that unlock entire teams.
• Designing systems that remain reliable even under messy real-world conditions.
• Figuring out how complex ML workflows should be orchestrated.
• Creating agentic behaviours that are explainable, debuggable, and self-healing.
• Making AI systems usable in products — not just demos.

I enjoy the craft of designing abstractions: choosing what’s exposed, what’s hidden, and how to create interfaces that make correctness the default.

I’m happiest when the work sits at the intersection of AI, systems design, and engineering tooling.

Technical Journey

My technical path is broad, but there’s a consistent theme:

I build systems and frameworks that make other systems — and other engineers — more capable.

• At Wayve, I work on the data/ML platform for autonomous vehicles: designing ingestion and transcoding frameworks for high-volume sensor data, and building the tooling that powers downstream ML and research workflows.
• At GSK, I designed a self-healing multi-agent system for automated exploratory data analysis, combining LLMs, orchestration logic, and evaluation loops as part of my AI master’s programme at KU Leuven.
• Before that, I built CI/CD frameworks, automated testing systems, and networking/embedded software at Ekinops, contributing to router-level protocol debugging and testing automation in both C and Python.
• Earlier still, I worked on ML-powered literature classification and electromagnetic simulations for Toyota’s Advanced Powertrain department.
• And finally, some PCB and electronics work, way back.

Now, my focus is clear:

(LLM) systems, agentic workflows, evaluation harnesses, and the tooling required to make AI dependable in production.

Get in Touch

I’m always open to interesting conversations about AI, software engineering, or collaboration opportunities. Feel free to reach out via LinkedIn or email.