From Dublin
to the world.
Since 2017.
We're a computer vision and AI company. We build systems that see something in the physical world, judge it, and turn that judgement into action — on a production line, an offshore platform, a showroom floor or a film set.
Seeing, deciding, instructing — applied where it earns its keep.
See
Detecting, recognising and measuring what a camera captures — defects, objects, contexts, people, environments.
Decide
Turning perception into a judgement: pass or fail, this part, this step, this configuration, this risk.
Instruct
Delivering that decision where it's needed — as AR on the object, a signal to a machine, or a record in your systems.
We've applied it in manufacturing and quality control, oil & gas and construction sites, automotive sales, and entertainment. The domain changes; the engineering underneath does not.
The work happens on their site, not in our office.
Four steps, and none of them start with a slide deck.
We start on your site
One visit, one problem, real conditions. Lighting, dust, cycle time, the people who'll actually use it — all of it is part of the spec.
We prove it before we build it
A narrow pilot on your own data and your own objects. If the vision doesn't hold up in your environment, you find out in weeks, not quarters.
We engineer for the shift, not the demo
Models, hardware, AR, and integration into the systems you already run. It has to survive Monday morning without us in the room.
We stay after go-live
Drift, new part numbers, new sites. Visual systems degrade quietly; ours are monitored and retrained.
Ancient instinct, modern engineering.
Prediction is worthless until it changes an action.
The mechanism from Antikythera didn't just describe the sky — it told people when to plant, sail and gather. Every system we ship is judged the same way: did someone do something different because of it?
The hard part is the last metre.
Getting a model to 95% is a known problem. Getting the answer onto the object, in front of the right person, at the right second, in a place with bad light and no signal — that's the engineering.
New ways of thinking beat more compute.
Most of our breakthroughs came from reframing the question, not from a bigger model. We invent around constraints because our customers live inside them.
Two people who kept asking what a machine could see.
Paula Guimarães
Kristian Karazissis
Two thousand years, one direction.
Off Antikythera, someone builds a bronze machine that reads the sky and hands back a usable answer. We take the name, and the brief.
Founded in Dublin.
€5M raised. Team scales across CV, AI, AR, and systems.
Lynx general availability — autonomous visual QC on the line.
Antikytera consolidates around one capability: applied visual intelligence. Ermes launches as flagship.
New funding round in progress.
AI specialists · computer vision engineers · AR developers · systems architects · UX/UI designers · trusted advisors.
Enterprise Ireland · NOF · Waterstart · Futurology.
Bring us something you need a machine to see.
A line, a site, a showroom, a set. If a camera can look at it, there's usually a decision hiding in it.