Skip to main content

Platform

The stack we start from, and the reasoning behind it.

This is not a product you license. It is the reference architecture and agent runtime our engagements converge on, published so you can argue with it before you hire us.

On-premises reference architectureSix layers inside the client security boundary. From the bottom: infrastructure (GPU nodes, block and object storage, network boundary); serving (inference engines, model router, KV and prompt cache); data and retrieval (vector index, document store, connectors); agent runtime (planner, tool broker, memory, policy engine, state store); governance (identity, authorisation, audit log, evaluation, observability), which spans the full width; and interfaces (APIs, internal UI, existing line-of-business systems). Nothing crosses the boundary outward.client security boundary, no egressINTERFACESInternal APIsOIDC-protectedOperator consoletraces, approvalsLine-of-business systemsITSM, CRM, EDRMBatch / scheduled jobsAGENT RUNTIMEPlannergraph executionTool brokertyped contracts, MCPPolicy engineleast privilegeMemoryworking / episodicState storecheckpoint + replayDATA & RETRIEVALVector indexhybrid + rerankDocument storeACL-awareFeature storeSource connectorsSERVINGInference enginevLLM / SGLang / NIMModel routersize-to-taskPrompt + KV cacheEmbedding + rerankINFRASTRUCTUREGPU nodes (Kubernetes, MIG-aware)Block + object storageSegmented network, no outbound routeGovernance plane · identity (AD/LDAP, OIDC) · authorisation · immutable audit log · evaluation · OpenTelemetry
Figure 1. The reference stack. The governance plane spans every layer because retrofitting it to one layer at a time does not work.

Architecture

Six layers inside one boundary, what sits in each, and what we would use instead under different constraints.

Architecture

Deployment topologies

On-premises, private cloud, sovereign, air-gapped and edge, compared on egress, identity, update path and operational burden.

Deployment topologies

Evaluation

How we measure agents that have no single right answer, and what an evaluation report contains.

Evaluation

Security model

Threat model, isolation, prompt-injection mitigations, supply-chain integrity and logging.

Security model

Start with the constraint.

Most of these projects are shaped by what you cannot do rather than what you want. Data that cannot leave the estate, a model you cannot host with a third party, a decision somebody has to justify to a regulator. Tell us yours and we will say honestly whether we can work inside it.

Talk to an engineer