HackYeah 2026 ReviewCouncil score
Median of 3 models, weighted by the task's official criteriaA well-engineered, deployed apprenticeship marketplace that fails the ImpactHer brief, since nothing beyond a feminine name addresses a challenge disproportionately affecting women, and the submission arrives without a deck, tests, or a stated problem.
A two-sided craft-apprenticeship marketplace with sensible domain details (exam goal, distance filters, learning forms), but the pattern is familiar and no need disproportionately affecting women is identified or evidenced. The blank problem-statement field supports the lower reading shared by B and C over A's 4.
Members disagree here: scores range by 2.0 points.Nothing in the product only makes sense for women; the sole tie is the feminine name, and the submission names no women-specific need or impact plan. The evidence better supports the near-zero view of A and B than C's 3, since a name is not a category connection, and the missed category requirement is listed among the weaknesses.
Members disagree here: scores range by 3.0 points.The app works end to end on a live demo with a mobile-first Polish UI and demo accounts, so it is usable as a generic product. But it serves no identified group under this task's theme, and with no deck, a largely blank form, and no impact plan, the median lands between A's generous 6 and B's strict 3.
Members disagree here: scores range by 3.0 points.The code shows deliberate, cohesive craft: a hand-drawn SVG hero, consistent color tokens, the Zilla Slab and Manrope pairing, 44px touch targets, and aria labels. C's 5 is the outlier against two 7s, though zero screenshots leave the rendered result unverified.
Members disagree here: scores range by 2.0 points.The full loop is deployed: auth, haversine search RPCs, a request status machine, reviews gated by accepted apprenticeships in RLS, migrations, and rich seed data. Zero tests, no deck, and one post-deadline commit keep it below B's 8.
Members disagree here: scores range by 2.0 points.Strengths
- A genuinely working two-sided marketplace, live at czeladniczka.vercel.app (HTTP 200): Supabase auth, distance-filtered search RPCs, a request and invite lifecycle, and reviews gated by accepted apprenticeships in database policy.
- Careful data layer: migrations with check constraints, RLS on every table, column-level grants, a separate hardening migration, and realistic seed data including review history.
- Polished mobile-first Polish UI with accessibility attention (44px targets, aria labels) and a README with demo accounts so judges can run it.
- A real weekend build: 27 of 28 commits during the event by 2 authors, about 1,945 lines of source across 33 files.
Weaknesses
- Misses the task's core requirement: no women-specific need, feature, evidence, or measurable-impact plan exists anywhere in the submission; stripped of its feminine name the product is entirely gender-neutral.
- Missing deliverables: no deck, no screenshots, and the submission form's problem and goal fields are essentially empty, so there is no problem statement or impact case.
- Zero tests (0 files, 0 cases) despite a database-heavy design with policies and RPCs that need them.
- No validation that the matching helps anyone: no target group, no impact metrics, and no sign of user research.
Red flags
- May not be built for this task: the project is a gender-neutral craft-apprenticeship matching platform whose only tie to ImpactHer is the feminine Polish word in its name; no women-specific need, feature, evidence, or impact plan appears in the form, README, code, or demo.
- One commit landed after the Sunday 11:00 deadline (minor, but a measured fact).
claude:glm-5.3-flash47.580% agreedots-studio/dots-3-note-preview:free34.090% agreeinclusionai/ling-3.0-flash-sante:free36.070% agreeclaude:glm-5.3–judgeThe council, the evidence pack, the prompts and the queue are all on GitHub. If a review helped you, a star helps other teams find it.