Rta Labs
Ṛta ऋत
In Vedic philosophy, Ṛta is the order that governs the cosmos, expressed in natural law and truth.
Rta Labs takes its name from this principle. Across its three divisions, the starting question is how the parts of a system work together. An agent needs context from the people it serves. A GPU deployment depends on power, space and capital. A model’s behavior depends on how it carries information from one step to the next.
Our work
Modular AI agents
Deploy a preconfigured Lattice or build custom agents around your tools and data. Specialized agents share context as they hand work from one step to the next. Start with one complex workflow, then connect agents across departments as your operations grow.
Explore OMNIAI infrastructure
From a first GPU cluster to a data center expansion, Compute helps organizations plan and fund AI deployments. Its global partner network gives buyers a choice of GPU capacity, hardware, powered colocation and connectivity. Advisors coordinate sourcing and help arrange equipment financing.
Talk to ComputeFrontier research
Frontier explores new model architectures and ways to understand their behavior. Its current research centers on Gravitational Vector Networks (GVN).
Read about GVNGravitational
Vector Networks
When a model reads a sequence, what does it carry forward from one step to the next?
GVN takes inspiration from geometry and dynamical systems. The research asks whether tracing changes in a model’s internal state can reveal how it uses earlier information as it processes a sequence.
Pirassena Sabaratnam
Pirassena founded two startups before Rta Labs and has spent more than a decade working in AI and enterprise software.
Connect on LinkedIn ↗Follow the work
Get product launches and research milestones by email, or read the latest OMNI news.