Responsible portfolio AI

AI trust centre

The portfolio AI is locally developed, self-hosted, grounded in approved public evidence, and designed to show limitations instead of making hiring decisions.

Privacy and grounding

The assistant retrieves only approved portfolio summaries and public repository metadata. Role descriptions are processed transiently and are not added to the knowledge base.

Review boundaries

Answers expose their sources, comparison claims have stable evidence IDs, and evidence gaps remain explicit. The assistant cannot verify private records or produce a hiring score.

Open the evidence workbench.