The concept
Every question is generated from the transcript of a real CBT Nuggets video and tied to a specific testable concept — not a lesson title, not a generic question bank. Each one carries its source citation forward, so an answer is always traceable to the passage that teaches it.
Learning theory
Every item is tagged to a level of Bloom's taxonomy and selected so an assessment covers a course's concepts and objectives at a deliberate cognitive mix — weighted toward reasoning, not rote. The same vetted pool powers a placement check before training and a mastery check after, so the before/after delta reflects measured change against the same validated content.
Learning technology
Each generated question runs a registry of validation gates — structural checks, a grounding gate that fails closed unless the answer is derivable from the cited material, distractor-quality classification, and subject-matter-expert sign-off. Item statistics from real answers then retire questions that stop discriminating.
How a question is made
The hard part — generate and validate grounded questions — is done once, into a reusable bank. Each assessment is then a selection over that bank for a purpose.
The platform walks every video's ordered transcript, embeds the passages, and clusters them into distinct testable concepts — a structured table of contents at the granularity of things a learner should walk away knowing.
For each concept, a model is given the concept and the transcript passage that teaches it, and writes a question that tests understanding of that material — citation attached.
Structural validity, factual grounding (fail-closed), reasoning-vs-recall, and distractor plausibility are checked. Hard gates eliminate; soft gates feed a quality score the selection layer ranks on.
Subject-matter experts review and sign off. The validation stack is the first line of defense; SME review and item-statistics-driven retirement are the second and third.
As learners answer, per-item statistics (difficulty and discrimination) accumulate. Items that stop telling who knows the material apart are flagged for retirement, and coverage gaps route back as signal.
Grounded in learning science
Bloom’s taxonomy is the standard framework learning scientists use to classify what a question actually demands of the test-taker — from recalling a fact to defending a decision. We tag every learning objective and every questionto a level of the revised 2001 taxonomy, then shape each assessment around that mix. It’s the difference between a quiz and a measurement.
Recall is the floor, not the goal. A score built on memorization overstates readiness — tagging every item to a cognitive level lets us measure whether your team can apply and analyze, the levels that predict on-the-job performance.
Every assessment is shaped to a deliberate Bloom mix, with a hard ceiling on pure-recall items and the weight pushed toward application and analysis. The number you see reflects capability, not cramming.
Placement and mastery draw from the same level-tagged bank, so the before/after delta measures real growth in reasoning — and the post-test tilts further toward apply, analyze, and evaluate to prove transfer, not retention.
We deliberately stop at Evaluate: the sixth level, Create, rewards producing original work — which a multiple-choice item can’t fairly measure, so we never claim to.
One bank, measured two ways
Because placement and mastery draw from the same validated pool, the delta between them measures real change. And because questions map to official certification blueprints, readiness rolls up by domain — for the learner and for the team lead deciding where to invest training.
A short, concept-spanning check that surfaces a per-skill picture of strengths and gaps — so a learner (or their manager) sees what to skip and what to focus on.
Questions tag back to official exam blueprints — domain weights and objectives — so readiness rolls up per domain and the weakest objectives surface for follow-up training.
Sign in to author and review the question bank, or to take your assessment.
