Model validation report
Model documentation and tests reviewed to the standard
How a run goes
The steps this agent takes, in order, and what each one leaves behind.
Reads the model documentation and repository
Reviews a model's documentation, code and test results against the validation standard
Runs the validation tests and compares results
Reruns the benchmark and sensitivity tests and records where the results differ from the developer's
Drafts the report with graded findings
Drafts the validation report with findings graded and the conditions for approval
Works in
Databricks, GitHub, Confluence
Databricks, GitHub, ConfluenceWhat it does
- Reviews a model's documentation, code and test results against the validation standard
- Reruns the benchmark and sensitivity tests and records where the results differ from the developer's
- Drafts the validation report with findings graded and the conditions for approval
Connects to
Need it to work differently?
Every agent can be changed: the steps it takes, the systems it uses and who approves what.