01 / PLAN
Architect
Turns intent, repository state and engineering rules into a bounded Builder Contract.
A Sentrium engineering project
SENTINEL
Engineering AI systems
that can show their work.
Built independently in regional NSW, Sentinel is being designed as a governed AI engineering platform focused on software systems that are structured, auditable and recoverable.
About
Software engineering with an evidence trail.
Sentinel is being built around a simple idea: AI engineering should not be a black box. Planning, implementation, review and evidence recording should be separated, governed and recoverable — so the system can explain what happened, why it happened and where human intervention was required.
Interactive preview
Click through the engineering spine.
This is a lightweight public preview of the architecture. Select a module to see the role it plays in a governed build.
02 / BUILD
Builder
Implements the contract with the smallest sufficient change and produces execution evidence.
03 / REVIEW
Reviewer
Checks the implementation independently against the contract, tests and engineering standard.
04 / EVIDENCE
Recorder
Preserves structured evidence so decisions, changes and outcomes can be reconstructed later.
Architect · Builder Contract
Defines objective, scope, assumptions, constraints, acceptance criteria and required verification before implementation begins.
Vision
Make AI-built software easier to trust.
The long-term goal is not simply faster code generation. It is a system that can coordinate engineering work, detect when something went wrong, preserve the evidence and improve the upstream process that caused the failure.
The standard stays constant. Governance scales to the work. Evidence remains available after the build is over.