Skip to main content

Part XV (b) - Capstone Evaluation Rubric

HARD TRUTH: COMPLETION IS NOT MASTERY

This rubric evaluates behavior and outcomes that distinguish top engineers from strong implementers.

Use it after each capstone and in quarterly progression reviews.

TECHNICAL JUDGMENT (25%)

Assess:

  • Tradeoff quality under constraints
  • Decision clarity and documentation quality
  • Boundary and interface design
  • Reversibility awareness

Scoring guide:

  • 1: pattern copying, weak rationale
  • 3: reasonable decisions, partial depth
  • 5: strong, defensible choices with explicit tradeoffs

RELIABILITY AND CORRECTNESS (25%)

Assess:

  • Failure handling quality
  • Data correctness guarantees
  • Observability coverage
  • Incident readiness and runbook quality

Scoring guide:

  • 1: optimistic design, weak safeguards
  • 3: baseline controls present
  • 5: resilient behavior proven by drills/tests/incidents

DELIVERY EXCELLENCE (20%)

Assess:

  • Scope shaping
  • Milestone predictability
  • Risk tracking discipline
  • Rollout and rollback quality

Scoring guide:

  • 1: reactive and unstable execution
  • 3: acceptable planning with occasional drift
  • 5: stable delivery with clear correction loops

COMMUNICATION AND LEADERSHIP (15%)

Assess:

  • Design communication clarity
  • Stakeholder alignment quality
  • Cross-team collaboration behavior
  • Mentorship and leverage impact

Scoring guide:

  • 1: communication creates ambiguity
  • 3: communication supports delivery
  • 5: communication increases team quality and speed

BUSINESS IMPACT (15%)

Assess:

  • User outcome movement
  • Performance or cost impact
  • Strategic relevance
  • Sustainability of outcomes

Scoring guide:

  • 1: no measurable impact
  • 3: moderate impact with weak longevity
  • 5: clear and durable impact tied to product goals

War-Story Mini-Case: Rubric Prevented a Bad Promotion Call​

Timeline:

  • Week 0: Candidate project demo scores highly on visual polish and narrative confidence.
  • Week 1: Rubric review reveals weak reliability evidence and no measurable user outcome movement.
  • Week 1 panel: Promotion decision paused pending evidence-backed reassessment.
  • Week 2: Reliability weight aligned with technical-judgment weight; evidence links required for each criterion.
  • Week 4: Candidate resubmits with measurable impact and improvement plan.

Key decisions:

  • Enforced weighted scoring discipline over impression-based evaluation.
  • Required evidence links to support every high-score claim.
  • Used low-scoring dimensions to drive targeted development, not blanket rejection.

Outcome:

  • Promotion decisions became fairer, clearer, and more development-focused.
  • Rubric reduced bias from presentation quality alone.

OUTPUT ARTIFACT

For each capstone evaluation, produce:

  • Completed score sheet with weighted total
  • Evidence links per criterion
  • Strengths summary
  • Improvement plan for two lowest criteria

Field rule: use evidence, not narrative confidence, as the basis for scoring.