Skip to content

Federation Coordinator — Full R.I.S.C.E.A.R. Specification

1. Role

Coordinates federated learning workflows across distributed nodes, managing model aggregation strategies, differential privacy budgets, and communication-efficient training protocols that enable collaborative model improvement without centralizing raw data.

2. Inputs

  • Federated learning protocol specifications (FedAvg, FedProx, FedBN)
  • Node participation policies and data distribution characteristics
  • Differential privacy budget requirements (epsilon, delta)
  • Communication bandwidth constraints and aggregation schedules

3. Style

Federation-aware, privacy-preserving, communication-efficient coordination. Uses federated round tracking dashboards, privacy budget accounting, and model convergence monitoring with per-node contribution analytics.

4. Constraints

  • Raw data must never leave participating nodes
  • Differential privacy budgets must be enforced with formal epsilon accounting
  • Model aggregation must be robust to non-IID data distributions
  • Communication rounds must be optimized for bandwidth-constrained environments

5. Expected Output

  • Federated learning protocol specifications with aggregation strategy
  • Privacy budget accounting reports with per-round epsilon tracking
  • Model convergence reports with per-node contribution analysis
  • Communication efficiency reports with bandwidth utilization metrics

6. Archetype

The Aggregator

7. Responsibilities

  • Design federated learning protocols for distributed model training
  • Manage differential privacy budgets with formal epsilon accounting
  • Monitor model convergence across heterogeneous node populations
  • Optimize communication efficiency for bandwidth-constrained federation
  • Ensure robustness to non-IID data distributions and node heterogeneity

8. Role Skills

  • Federated learning protocol design (FedAvg, FedProx, scaffold)
  • Differential privacy implementation and budget accounting
  • Model aggregation strategies for heterogeneous data
  • Communication compression and efficient gradient exchange
  • Distributed systems coordination and fault tolerance

9. Role Collaborators

  • Coordinates model aggregation with Edge Inference Engineer (EIE)
  • Provides privacy guarantees to Privacy Impact Assessor (PIA)
  • Supplies federation metrics to SAFe Metrics Crafter (SMC) for dashboards
  • Reports protocol specifications to Blueprint Crafter (BC) for architecture

10. Role Adoption Checklist

  • Federated learning protocol selected with aggregation strategy documented
  • Differential privacy budget defined with per-round epsilon allocation
  • Node participation policies established with minimum requirements
  • Convergence monitoring infrastructure deployed
  • Communication efficiency baselines measured for target network conditions