CrackIt Dev

Scores reflect how complete and interview-ready your RADIO submission is for practice. They help you prioritize gaps — they are not a prediction of how you will perform in a real loop or whether a company will hire you.

Overall score (0–100)

Your total is the sum of five rubric sections (max 100 points). The reviewer grades each section against a question-specific scoring guide, then maps the sum to a band label.

BandMinimum scoreWhat it means (practice)
Staff-like88+Strong scope, architecture, and tradeoffs — bar for staff-level practice depth, not a staff offer.
Strong75+Solid structure with minor gaps; good mock-interview readiness with targeted fixes.
Competent60+Core ideas present; needs sharper requirements, data model, or reliability detail.
Basic40+Partial coverage; several RADIO steps thin or missing key front-end concerns.
Weak0+Major sections missing or superficial; revisit the question and RADIO flow.

Rubric sections (backend)

The AI reviewer scores these five areas. Point caps come from REVIEW_SECTION_MAX.

  • Requirementsup to 15 pts
  • Architectureup to 25 pts
  • Sync & reliabilityup to 25 pts
  • Securityup to 20 pts
  • Tradeoffs & communicationup to 15 pts

Per-step RADIO breakdown (review UI)

On your review page, each RADIO step is shown on a 0–20 display scale. Raw section scores are mapped from the rubric using the same logic as reviewBreakdown: Architecture is split across A / D / I; Optimizations (O) blends sync/reliability and security; Tradeoffs maps directly. Empty or thin steps can cap the displayed bar.

StepRubric sourceSection maxReview UI
RRequirementsRequirements15Scaled to 0–20 bar
AArchitectureArchitecture (42% of section)11Scaled to 0–20 bar
DData modelArchitecture (33% of section)8Scaled to 0–20 bar
IInterface & syncArchitecture (25% of section)6Scaled to 0–20 bar
OOptimizationsSync & reliability + Security (averaged)23Scaled to 0–20 bar
TTradeoffsTradeoffs & communication15Scaled to 0–20 bar

AI & your data

Reviews are generated by an AI model using your submitted notes, canvas, and Mermaid for that session. We do not use your practice submissions to train third-party models. See our Privacy Policy for how we handle account data and AI features.