Vareon builds Duplex.
The runtime intelligence for your AI and your agents.
Duplex is a second intelligence beside your AI that learns and acts at runtime, for one person, a team or a whole enterprise. What it builds is yours, whichever model you use.
Duplex is a product of Vareon, Inc., Irvine, California
What Vareon means.
Var, for variation and change. Eon, for time. Vareon is change over time: adaptation.
The mark says it too. It’s ∇, the nabla: in mathematics, the direction of change, and the gradient every model learns by. Drawn as an impossible loop, each beam in front of the next, it has no beginning and no end. Change, over time.
- Var
Variation, change
Every result changes something: a fix, a rule, a better way to do the work.
- eon
Time
Not once, at a release, but for as long as the work goes on, each change building on the last.
- Vareon
Change over time: adaptation
What a trained model can’t do once it ships, and what ACI was built to do. Duplex brings it to your work.
We created ACI.
ACI, Adaptive Continual Intelligence, is a new category of intelligence that works in tandem with AI models. Duplex is built on it.
Step 1 of 4: Who could
Until now, only the labs that train models could make AI learn. With ACI, every person and company can.
One founder, a new kind of intelligence.
Faruk Guney built ACI, Duplex and his new startup on his own.
For more than fifteen years he grew Norck from a California startup into a global on-demand manufacturing and engineering company, headquartered in Germany. He saw where AI falls short in real work: it learns once, before release, and can’t learn and act on the job.
So he spent years on the research question at its root: the trilemma of neural nets, plasticity, stability and editability. ACI is the answer, and Duplex brings it to everyday work.
He believes in principled acceleration: AI as the engine that speeds up progress in materials, energy and engineering, with every step traceable, trusted and true to how the real world works.
Faruk Guney
Founder · AI and industrial engineering
From scaling global manufacturing at Norck to engineering intelligence for the physical world, he builds systems for the problems where physics matters and the margin for error is zero.
2025 to today
Founder
ACI, Duplex and a new startup building AI agents and robotic systems that learn continually and act in the physical world.
2009 to 2025
CEO and founder, Norck
Prototype. Manufacture. Scale up. On demand. From garage startups to Fortune 500 companies, worldwide.
Before Norck
Engineering
Logistics systems engineering lead at Anna’s Linens, manufacturing engineer at Zodiac Aerospace Services, and productivity analyst for the republic of Türkiye ministry of industry and technology.
Education
USC and Johns Hopkins
M.s. In industrial and systems engineering at USC, and advanced studies in artificial intelligence at Johns Hopkins university.
Agents will get better with a second intelligence beside them, in a recursive loop, at runtime.
Recursive self-improvement will emerge at runtime. It is a significant departure from established ideas, and we believe it is the future of AI.
Step 1 of 5: An agent works
It will come while the work happens, from a second intelligence beside every agent, rather than from coding agents editing their own code or from weights retrained on clusters of accelerators.
What we believe.
Three beliefs, and everything we build follows from them.
- 01Every AI needs a second intelligence beside it.
- 02What it learns belongs to you.
- 03Your whole stack should learn together.
Step 1 of 3: A second intelligence
Your AI learned the world slowly. Duplex learns your work fast and acts on it: two intelligences, one system.
Why we built Duplex.
Your AI learned the world before you arrived. Your work needs an intelligence that learns and acts while it happens, so we built one.
Step 1 of 4: Before you arrive
Your AI learned the world slowly, long before it met your work.
One product, five integrations, three places to run it.
Duplex works inside Claude, Codex, Cursor and Hermes, and inside your own agents through MCP + hooks, in Vareon Cloud, your cloud or no cloud.
Step 1 of 3: Your assistants
In your assistants, Duplex works for you and your team: inside Claude, Codex, Cursor and Hermes.
Results on real work.
On the same work and the same model, Duplex was more accurate, cost a fraction as much and stayed right when the rules changed.
Step 1 of 5: Accuracy
On 3,000 decisions, Duplex beside the model reached 94.6% accuracy, against 59.8% for the model alone.
Tell us what Duplex should learn first.
You start with a pilot on your own work, with the measures agreed before it begins. Or write to hello@vareon.com.