What runs underneath the r* products: models applied to specific product decisions, in-house GPU compute, real-time data engineering, and APIs with clear boundaries. Each one is verifiable in the products that use it.
The shared foundation makes models available to every product; a product uses them where it needs a decision made — an answer, a score, a segment, an action. The three applications below are where that is verifiable today.
Answering from your knowledge, choosing the next step and completing it across systems — the engine behind rAgent.
Linking trading, account and payment signals into explained scores — the engine behind rMIB.
Which segment, which communication, which limit — inside rPAM’s account operations.
RoboAI operates its own GPU compute for model development and inference, shared across product teams. One foundation serves seven products; no product has to build its own.
Training and evaluation for the product models above.
Serving models inside the products, close to the data they act on.
Integrating external providers, normalising streams, running high-load services and keeping operational history — the discipline rMarket and rGateway are built on.
A product is deployed, configured and billed on its own. Where two r* products are connected, the integration is native and implemented — not every pairing exists. Towards your systems, integration goes through documented APIs, and the connectors that are live for your case are disclosed in the demo flow.
Accounts, payments, risk, execution and experience meet without custom work on the connections that exist — each product page lists its own.
APIs for the systems you already run; third-party connectors stated per product.