Enterprise software reality
Buyers need controls, logs, rollback paths, and repeatable deployment before a model becomes a workflow.
Founder
Operator-led AI for environments that cannot gamble on cloud.
Maxime Tolos combines enterprise systems work with two decades of hospitality operations. The product bias is practical: deploy locally, prove the workflow, and leave operators with controls they can use.
Why this wedge is believable
The hard part is not a model. It is deployment people can trust.
The founder-market fit is at the intersection of enterprise systems, local operations, and measured execution. That is exactly where cloud-blocked AI projects fail.
Service reliability mindset
Hotel operations trains a bias toward uptime, escalation, handoff, and clear evidence over theory.
Lean beta discipline
Current proof stays narrow until the sanitized demo and benchmark harness can support stronger claims.
Operating principles
How the product decisions are made.
Sovereign AI Platform is intentionally scoped around the blocked buyer: teams with sensitive data, local infrastructure, and operators who must understand what is running.
Privacy
Privacy is a product feature
Local deployment and air-gapped operation are product requirements, not only legal language.
Operability
Controls before scale
Model/runtime state, RBAC, audit logs, offline updates, and diagnostics come before broad rollout.
Proof
Benchmarks before claims
Alpha 1 is measured against fixed quality and throughput targets using non-sensitive eval inputs.
Direct contact
Talk through the beta, investor memo, or design-partner fit.
Email Maxime at [email protected]. Include product, investor, or design partner in the subject.