Assessment & Grading
The Applied AI Fellowship uses artifact-based assessment focused on real, production-grade deliverables rather than exams or theoretical knowledge.
Assessment Philosophy
Artifact-Driven Learning
- Build real, reusable deliverables
- Demonstrate practical capability
- Create enterprise-grade artifacts
- Focus on production readiness
Universal Scoring Dimensions
All assessments evaluate five core dimensions:
Judgment
Quality of decisions under uncertainty
- Makes sound decisions with incomplete information
- Demonstrates restraint and tradeoff awareness
- Knows when to avoid AI solutions
- Balances risk and opportunity appropriately
Key Indicators:
- Explicitly identifies and defers unsuitable use cases
- Shows awareness of limitations and constraints
- Makes principled decisions under uncertainty
Rigor
Evidence-based reasoning
- Uses data and evidence to support decisions
- Applies proper evaluation methodologies
- Validates assumptions and hypotheses
- Demonstrates systematic thinking
Key Indicators:
- Evidence-based approach throughout
- Proper testing and validation
- Data-driven decision making
- Clear reasoning and justification
Transfer
Ability to apply learning to new contexts
- Adapts knowledge to new situations
- Generalizes from specific examples
- Applies frameworks across domains
- Demonstrates practical application
Key Indicators:
- Applies learning to novel contexts
- Transfers knowledge effectively
- Adapts frameworks appropriately
- Shows practical problem-solving
Governance
Awareness of risk, controls, and accountability
- Identifies and mitigates risks
- Implements proper controls and guardrails
- Ensures accountability and auditability
- Demonstrates governance awareness
Key Indicators:
- Risk assessment and mitigation
- Proper controls and guardrails
- Audit trails and documentation
- Accountability mechanisms
Tool & AI Mediation
Understanding how AI tools augment human capability
- Understands how AI tools shape thinking and decision-making
- Knows when to use AI tools vs. traditional methods
- Evaluates tool capabilities and limitations effectively
- Integrates AI tools into workflows appropriately
- Recognizes implications of tool-mediated work for governance
Key Indicators:
- Appropriate tool selection and use
- Understanding of tool limitations and constraints
- Effective integration of tool outputs with human judgment
- Awareness of how tools mediate cognition and capability
- Tool-mediated governance and accountability considerations
Pass Criteria
To successfully complete a track, you must meet:
- Minimum 70% overall - Across all five dimensions combined
- No critical governance failures - Governance is non-negotiable; critical failures result in automatic failure regardless of other scores
- All required artifacts submitted - Including your track’s capstone package as specified below (see Credentialing for how this maps to badges)
Distinction Criteria
For exceptional performance, demonstrate:
- Restraint and tradeoff awareness - Shows judgment in when not to use AI
- Explicitly kills or defers unsuitable AI use cases - Demonstrates good judgment
- Produces reusable, enterprise-grade artifacts - Production quality deliverables
Artifact Requirements by Track
Each track builds through weekly work and learning levels toward one credentialed capstone artifact for that track. Cohorts may require interim drafts or checkpoints; what gets scored for the badge is the capstone unless your intake says otherwise.
T4 vs T5: Both involve architecture, but the emphasis differs. T4 (AI Engineer) is system-first: a production-grade agentic system plus documentation sufficient to operate, secure, and observe it. T5 (AI Architect) is design-first: a production-quality architecture document and annotated stack diagram, with depth across planes (memory, data, integrations, security, observability). T5 work may reference or extend T4-style delivery patterns; it does not substitute for T4’s requirement to ship and document a runnable system.
T1: Executive
Required Artifact: AI Investment Decision Framework
- Investment thesis and prioritization logic inside the framework
- Risk assessment suited to executive and board review
- Governance and accountability structure
- Repeatable decision criteria (for example: fund, defer, or stop)
T2: AI Product
Required Artifact: Agentic Product Roadmap + Pricing Canvas
- Product strategy document
- Value proposition design
- Roadmap with milestones and pricing logic
- Use case prioritization framework
T3: Delivery Leader
Required Artifact: Governed AI Delivery Plan + Risk Register
- Delivery plan with governance checkpoints
- Risk register tied to planes and milestones
- Observability and runtime considerations
- Progressive delivery approach
T4: AI Engineer
Required Artifact: Production agentic system with documented architecture
- Production-grade agentic system
- Guardrails, identity, and control-plane integration
- Documentation and architecture
- Testing, observability, and validation results
T5: AI Architect
Required Artifact: Production architecture doc + annotated stack diagram
- End-to-end architecture across relevant planes
- Memory, data, and context design
- Security and observability integration
- Intelligence boundary controls
Assessment Process
- Artifact Submission - Submit required artifacts for your track
- Panel Review - Expert panel reviews artifacts across all dimensions
- Scoring - Each dimension scored independently
- Feedback - Detailed feedback provided on strengths and areas for improvement
- Credentialing - Badge issued upon successful completion
Public Sector Assessment
Public sector assessments include additional emphasis on:
- Accountability - Clear responsibility and ownership
- Auditability - Full traceability and documentation
- Explainability - Ability to explain system behavior
These align with regulatory requirements including PDPA, GDPR, and AI governance frameworks.
Ready to begin your assessment journey? Get in touch to learn more or apply.
Previous: Credentialing & Badges | Next: Public Sector Variant