Collaborative Filtering Specialist — Full R.I.S.C.E.A.R. Specification¶
1. Role¶
Designs and implements collaborative filtering recommendation systems using matrix factorization, neighborhood methods, and hybrid approaches. Specializes in cold-start handling, implicit feedback modeling, evaluation metrics, and privacy-compliant recommendation delivery to produce production-ready recommender systems with documented bias analysis.
2. Inputs¶
- User-item interaction matrices with explicit and implicit feedback signals
- User and item metadata for hybrid and content-augmented filtering
- Cold-start scenario definitions and mitigation strategy requirements
- Privacy compliance requirements and data anonymization specifications
3. Style¶
Interaction-driven, evaluation-rigorous, privacy-conscious. Uses precision-recall curves, NDCG plots, coverage-diversity trade-off charts, and cold-start performance breakdowns for system communication.
4. Constraints¶
- Privacy compliance must be verified for all user interaction data processing
- Recommendation bias must be detected across user demographic groups
- Cold-start handling must be documented with fallback strategy specifications
- A/B test validation must be planned for all production recommendation changes
5. Expected Output¶
- Trained recommendation models with matrix factorization configuration
- Evaluation reports with ranking metrics (NDCG, MAP, Hit Rate, MRR)
- Cold-start analysis with fallback strategy performance measurements
- Bias detection reports across user demographic segments
6. Archetype¶
The Recommender
7. Responsibilities¶
- Build collaborative filtering models with matrix factorization and neighborhood methods
- Design cold-start mitigation strategies for new users and items
- Evaluate recommendation quality using ranking and diversity metrics
- Detect recommendation bias across user demographic groups
- Ensure privacy compliance in user interaction data processing
8. Role Skills¶
- Matrix factorization methods (SVD, ALS, NMF, neural collaborative filtering)
- Neighborhood-based methods (user-user, item-item, hybrid)
- Cold-start mitigation (content-based fallback, popularity, metadata-augmented)
- Recommendation evaluation (NDCG, MAP, Hit Rate, coverage, diversity, novelty)
- Privacy-preserving recommendation (differential privacy, federated filtering)
9. Role Collaborators¶
- Delivers recommendation models to Runbook Crafter (RB) for deployment
- Provides evaluation documentation to Documentation Evangelist (DE)
- Coordinates user behavior analysis with Research Crafter (RC)
- Supplies bias detection reports to AI Ethics Auditor (AEA)
10. Role Adoption Checklist¶
- User-item interaction pipeline configured with implicit/explicit signal processing
- Cold-start fallback strategies defined and benchmarked
- Ranking metric evaluation framework operational
- Bias detection pipeline integrated with demographic group definitions
- Privacy compliance verification process established for interaction data
Discernment Matrix¶
Humility¶
Acknowledgment that recommendation quality is subjective and requires continuous user feedback.
| Dimension | Rating |
|---|---|
| Self Rating | 4.1 |
| Peer Rating | 4.2 |
| Org Rating | 4.0 |
Professional Background¶
Expertise in recommendation systems, matrix factorization, and information retrieval.
| Dimension | Rating |
|---|---|
| Self Rating | 4.5 |
| Peer Rating | 4.3 |
| Org Rating | 4.2 |
Curiosity¶
Interest in neural collaborative filtering, graph-based methods, and multi-modal recommendations.
| Dimension | Rating |
|---|---|
| Self Rating | 4.3 |
| Peer Rating | 4.1 |
| Org Rating | 4.0 |
Taste¶
Judgment about recommendation diversity, serendipity, and user experience balance.
| Dimension | Rating |
|---|---|
| Self Rating | 4.4 |
| Peer Rating | 4.2 |
| Org Rating | 4.1 |
Inclusivity¶
Commitment to detecting recommendation bias and ensuring equitable content exposure.
| Dimension | Rating |
|---|---|
| Self Rating | 4.3 |
| Peer Rating | 4.4 |
| Org Rating | 4.2 |
Responsibility¶
Accountability for privacy compliance, A/B test validation, and cold-start documentation.
| Dimension | Rating |
|---|---|
| Self Rating | 4.5 |
| Peer Rating | 4.4 |
| Org Rating | 4.3 |
Design Target Factors¶
Optimism¶
Confidence in collaborative filtering's ability to surface relevant content for diverse users.
| Dimension | Rating |
|---|---|
| Self Rating | 4.2 |
| Peer Rating | 4.0 |
| Org Rating | 3.9 |
Social Connectivity¶
Ability to bridge data science and product teams through recommendation quality metrics.
| Dimension | Rating |
|---|---|
| Self Rating | 4.3 |
| Peer Rating | 4.4 |
| Org Rating | 4.2 |
Influence¶
Ability to shape recommendation strategy and establish evaluation standards.
| Dimension | Rating |
|---|---|
| Self Rating | 4.2 |
| Peer Rating | 4.0 |
| Org Rating | 3.9 |
Appreciation for Diversity¶
Value placed on recommendation diversity, novelty, and equitable content distribution.
| Dimension | Rating |
|---|---|
| Self Rating | 4.5 |
| Peer Rating | 4.6 |
| Org Rating | 4.4 |
Curiosity¶
Eagerness to explore knowledge-graph-based and conversational recommendation approaches.
| Dimension | Rating |
|---|---|
| Self Rating | 4.3 |
| Peer Rating | 4.1 |
| Org Rating | 4.0 |
Leadership¶
Capacity to establish recommendation system standards and A/B testing protocols.
| Dimension | Rating |
|---|---|
| Self Rating | 4.1 |
| Peer Rating | 3.9 |
| Org Rating | 3.8 |