Company

Greek AI that Greek businesses can control, audit and afford.

Hellas Inference is an Athens company that runs Greek business workloads on Greek infrastructure. We are not a model lab and not a consultancy: we operate the products and take responsibility for the outcome.

Mission

Make Greek AI something a Greek business can control and call its own

Greek enterprises pay twice for AI today: for integration projects that cannot be reused, and for open-ended token bills to US model vendors. We replace both with outcome-priced products on an open-weight stack that keeps data in Greece and is auditable by construction.

Greek first

Models fine-tuned on Greek speech and text from real workloads, and judged on Greek data.

Sovereign by construction

Data and inference in Greece, open weights, and an immutable audit trail for every decision.

Priced on outcomes

You pay for calls handled, bookings made and answers delivered, never metered per token.

Measured, not claimed

We publish measurements from our own GPUs, weak spots included.

Founder

Leadership

John Doxaras, Founder and CEO of Hellas Inference

John Doxaras

Founder & CEO

John Doxaras built the Krikri and Gemma inference programme behind our benchmarks, a real-time Greek voice stack and the AI Employee platform. The founding shareholder of Hellas Inference is TRYROSO Μονοπρόσωπη Ι.Κ.Ε.

We are hiring

Join the team in Athens

We are hiring a CTO / ML lead and Greek-speaking ML engineers. If you have shipped speech or language models to production and want to do it for Greece, write to us.

Where we are

Athens, Greece

Our inference stack runs in production on contact-centre workloads for a Greek telecom operator and a Greek bank; the AI Employee platform is live at a leading Greek publisher. See the measurements

[email protected]

Legal status

Hellas Inference is under incorporation by TRYROSO Μονοπρόσωπη Ι.Κ.Ε. (ΓΕΜΗ 146333001000, ΑΦΜ 800985566), Athens. Until incorporation completes, TRYROSO Μονοπρόσωπη Ι.Κ.Ε. is the contracting entity.

Start with one workload.

Tell us the calls, documents or questions you want handled. We measure it on your data before you commit.