Blog

Two rubrics, both build-in-public: how the grading engine works, and what interview loops actually test — sourced from evidence, not vibes.

• July 10, 2026how we grade

  How we grade: inside an AI interviewer that has to prove itself

  Rubrics with weights, golden answers, band calibration, and a red-flag rule. The machinery that keeps an AI grader honest, documented in public.

• July 10, 2026evidence

  What senior frontend interviews actually test in 2026

  We rebuilt our syllabus from roadmap.sh, the major interview handbooks, and real 2025–26 loop reports. This is what all three sources agree interviewers test.