We believe radio networks deserve their own intelligence architecture
The RA-NN is that architecture — one shared representation of the network that every control task builds on. When the system learns something about coverage, interference benefits. When mobility improves, load balancing inherits the context. Nothing starts from zero.
What we believe
Architecture before algorithms
The breakthrough isn't a better model — it's a better representation of the network. Get the architecture right and everything built on it learns faster.
Observe at the edge
Individual user measurements capture what network-level summaries erase. That's where the real behavior lives — and where the RA-NN starts.
Build for operators, not conferences
We ship what works on real networks, not what demos well on slides. Every design decision is tested against operational reality.
Telecom needs its own foundation
Radio networks have unique physical structure. They need a purpose-built intelligence architecture — not language models, not generic ML.
Our founders
Terence Li
Co-Founder & CEO
Terence spent two decades inside Tier 1 networks — from GSM to 5G. He led engineering teams through multiple generations of network evolution, seeing firsthand the limitations the RA-NN is designed to overcome.
Michael Chiaramonte
Co-Founder & CTO
Michael built AI-driven network automation at a Tier 1 carrier and kept hitting the same wall: legacy architectures that made real intelligence impossible. He designed the RA-NN as the architecture those projects needed but never had.
We're looking for design partners
Whether you operate networks, build infrastructure, or research the future of wireless — we'd like to talk.