Skip to content

Persona Comparison Guide

This guide helps you choose the right persona (or set of personas) for a given task. It provides use-case comparison matrices for the most commonly confused persona groups, a champion selection guide, and a text-based decision flowchart.

Core Personas: RC vs BC vs DE

The three foundational personas define the canonical Find-Create-Critique cycle. They are complementary, not interchangeable.

Dimension RC (Research Crafter) BC (Blueprint Crafter) DE (Documentation Evangelist)
FCC Phase Find Create Critique
Archetype The Investigator The Architect The Guardian
Primary Output Capability matrices, research inventories, traceability matrices Blueprints, API specs, data models, workflow definitions Quality reviews, style enforcement, publish decisions
Key Skill Synthesis and organization of information Translation of research into design Quality evaluation and standards enforcement
Input Source Source code, architecture docs, stakeholder feedback RC's research package BC's blueprints and design artifacts
Works Upstream Of BC DE RB, UG (via publishing handoff)
Works Downstream Of (project brief) RC BC
When to Use Starting a new project, evaluating capabilities, literature review Designing systems, writing specifications, defining APIs Reviewing deliverables, enforcing standards, gatekeeping releases

Decision rule: If you need to gather information, use RC. If you need to design something, use BC. If you need to evaluate quality, use DE.

ML Lifecycle: POR vs FAR vs ESC vs IOR

Four ML personas that are often confused because they all touch the model development pipeline.

Dimension POR (Pipeline Orchestrator) FAR (Feature Architect) ESC (Experiment Scientist) IOR (Inference Optimizer)
Category data_engineering ml_lifecycle ml_lifecycle ml_lifecycle
FCC Phase Build Create Create Build
Archetype The Conductor The Feature Engineer The Experimentalist The Optimizer
Primary Output Pipeline DAG definitions, orchestration configs Feature store schemas, transformation pipelines Experiment logs, hyperparameter reports Quantized models, distilled architectures
Focus Workflow orchestration and scheduling Feature engineering and storage Experiment tracking and comparison Production inference performance
When to Use Designing the overall pipeline topology Creating and managing feature transformations Running and comparing experiments Optimizing a trained model for deployment

Decision rule: POR orchestrates the workflow. FAR builds the features. ESC runs the experiments. IOR makes the result fast.

ML Models: NNS vs GBT vs RFS

Three model specialist personas covering different algorithm families.

Dimension NNS (Neural Network Specialist) GBT (Gradient Boosting Trainer) RFS (Random Forest Specialist)
Category ml_models ml_models ml_models
Archetype Deep Learning Engineer Gradient Ensemble Builder Forest Ensemble Builder
Model Types CNNs, RNNs, Transformers, GANs XGBoost, LightGBM, CatBoost Random Forest, Extra Trees
Best For Unstructured data (images, text, audio) Tabular data with feature importance Tabular data with noise robustness
Training Cost High (GPU/TPU required) Medium (CPU efficient) Low (highly parallelizable)
Interpretability Low (needs IRE for explanations) Medium (built-in feature importance) Medium (feature importance, partial dependence)
When to Use Deep learning tasks, sequence modeling Competitive modeling, Kaggle-style tabular Baseline modeling, ensemble foundations

Decision rule: Use NNS for unstructured data. Use GBT for competitive tabular performance. Use RFS for robust baselines and when interpretability matters.

Governance: DGS vs GCA vs PTE

Three governance personas with distinct scopes.

Dimension DGS (Data Governance Specialist) GCA (Governance Compliance Auditor) PTE (Privacy Taxonomy Engineer)
Category governance integration governance
FCC Phase All Critique Create
Archetype The Integrator The Auditor The Classifier
Scope Data flows, API contracts, service configurations Compliance audits, remediation plans, evidence logs Data classification, privacy regulation alignment
Primary Output Data governance policies, data flow diagrams Audit reports, remediation guides Privacy taxonomies, classification schemas
When to Use Establishing data governance foundations Auditing compliance against constitutions Classifying data sensitivity and privacy impact

Decision rule: DGS defines the rules. GCA audits against the rules. PTE classifies the data that the rules govern.

Champion Selection Guide

Champions are elevated personas that orchestrate teams of base personas. Choose the right champion based on the workflow phase you need to coordinate.

Champion Phase Orchestrates Use When...
RCHM (Research Crafter Champion) Find RC, CIA, STE, RIC You need to coordinate a multi-persona research effort
BCHM (Blueprint Crafter Champion) Create BC, BV, UMC, RIC You need to produce an integrated design package
UGCH (User Guide Crafter Champion) Create UG, SCP, EC You need to coordinate user-facing content publishing
RBCH (Runbook Crafter Champion) Create RB, GCA, TS You need to validate and publish operational documentation

Champion handoff chain: RCHM -> BCHM -> DE -> UGCH and BCHM -> RBCH

This chain ensures that research flows into design, design flows through quality review, and approved content flows into both user-facing and operational publishing.

Decision Flowchart

Use this text-based flowchart to select the right persona for a task.

START: What is your primary goal?
  |
  |-- Gather information? -----> Is it a research task?
  |                                |-- Yes --> RC (Research Crafter)
  |                                |-- Is it data sourcing? --> DSS (Data Sourcing Specialist)
  |                                |-- Is it indexing/cataloging? --> CIA (Catalog Indexer Architect)
  |
  |-- Design something? --------> Is it a blueprint/spec?
  |                                |-- Yes --> BC (Blueprint Crafter)
  |                                |-- Is it a feature pipeline? --> FAR (Feature Architect)
  |                                |-- Is it a UI mockup? --> UMC (UI Mockup Crafter)
  |                                |-- Is it a taxonomy? --> STE (Semantic Taxonomy Engineer)
  |
  |-- Evaluate quality? --------> Is it a document review?
  |                                |-- Yes --> DE (Documentation Evangelist)
  |                                |-- Is it a compliance audit? --> GCA (Governance Compliance Auditor)
  |                                |-- Is it a blueprint validation? --> BV (Blueprint Validator)
  |                                |-- Is it fact-checking? --> AMS (Anti-fact Mitigation Specialist)
  |
  |-- Build a model? -----------> What type of data?
  |                                |-- Unstructured (images, text) --> NNS (Neural Network Specialist)
  |                                |-- Tabular (performance) --> GBT (Gradient Boosting Trainer)
  |                                |-- Tabular (robustness) --> RFS (Random Forest Specialist)
  |
  |-- Deploy/operate? ----------> Is it CI/CD?
  |                                |-- Yes --> PBD (Pipeline Builder for DevOps)
  |                                |-- Is it model serving? --> MOS (Model Operations Specialist)
  |                                |-- Is it rollout strategy? --> JUS (Just-in-Time Update Specialist)
  |
  |-- Coordinate a team? -------> Which phase?
                                   |-- Find phase --> RCHM (Research Crafter Champion)
                                   |-- Create phase (design) --> BCHM (Blueprint Crafter Champion)
                                   |-- Create phase (user docs) --> UGCH (User Guide Crafter Champion)
                                   |-- Create phase (ops docs) --> RBCH (Runbook Crafter Champion)
                                   |-- Cross-team --> CO (Collaboration Orchestrator)

Category Quick Reference

All 20 persona categories at a glance.

Category Count Primary Phase Purpose
core 5 Find/Create/Critique Foundational FCC cycle
integration 8 Mixed Specialized capabilities
governance 4 All/Critique Data integrity and compliance
stakeholder 5 Mixed Cross-team coordination
champion 4 Orchestration Team coordination
data_engineering 6 Build Data pipeline construction
ml_lifecycle 9 Mixed End-to-end ML workflow
ml_models 11 Build Model-specific expertise
devops 3 Build/Critique CI/CD and deployment
app_development 2 Build Application construction
ux_visualization 6 Create/Build Design and visualization
protocol_engineering 6 Create/Build A2A/MCP protocol design
jv_governance 6 Mixed Joint venture oversight
open_science 4 Mixed FAIR and open access
responsible_ai 5 Critique AI ethics and fairness
jv_collaboration 4 Mixed Cross-org collaboration
docs_as_code 4 Create/Critique Documentation automation
privacy 3 Create/Critique Data privacy
knowledge_graph 3 Create/Build Ontology and KG construction
local_first_ai 4 Build Edge and local AI