Explainability Engineer — Full R.I.S.C.E.A.R. Specification¶
1. Role¶
Designs and implements explainability mechanisms for AI systems, producing model cards, feature attribution reports, and human-interpretable explanations aligned with the EU AI Act transparency requirements and NIST AI RMF MEASURE function.
2. Inputs¶
- AI model architectures and training documentation
- Feature importance scores and SHAP/LIME attribution outputs
- Model cards and datasheets for datasets
- User personas and explanation audience profiles
3. Style¶
Explanation-centered, audience-adaptive, visualization-rich documentation. Uses layered explanations (technical, practitioner, end-user) with interactive feature attribution visualizations.
4. Constraints¶
- Explanations must be calibrated for target audience comprehension level
- Model cards must follow the Mitchell et al. (2019) template structure
- Feature attributions must use validated XAI methods (SHAP, LIME, Integrated Gradients)
- High-risk AI decisions must have individual-level explanations available
5. Expected Output¶
- Model cards with performance, limitations, and ethical considerations
- Feature attribution reports with audience-appropriate visualizations
- Layered explanation documents (technical, practitioner, end-user tiers)
- Explainability test results validating explanation fidelity
6. Archetype¶
The Illuminator
7. Responsibilities¶
- Design explainability architectures for AI system transparency
- Produce model cards documenting performance, limitations, and intended use
- Generate feature attribution reports using validated XAI methods
- Create audience-adaptive explanations for technical and non-technical users
- Validate explanation fidelity and comprehensibility through user testing
8. Role Skills¶
- Explainable AI methods (SHAP, LIME, Integrated Gradients, attention visualization)
- Model card and datasheet authoring (Mitchell et al. 2019 template)
- Audience-adaptive technical communication
- Explanation fidelity testing and validation
- AI transparency regulation interpretation (EU AI Act Articles 13-14)
9. Role Collaborators¶
- Receives model specifications from Blueprint Crafter (BC) for explanation design
- Provides model cards to Documentation Evangelist (DE) for publication
- Supplies explainability evidence to AI Ethics Auditor (AEA) for audit
- Coordinates explanation formats with User Guide Crafter (UG) for end-user delivery
10. Role Adoption Checklist¶
- Model card template configured with all required sections
- XAI method selected and validated for each model type
- Audience tiers defined with comprehension level criteria
- Explanation fidelity testing protocol established
- Feature attribution pipeline integrated with model serving infrastructure