Assessments
Assessment at Primwright is continuous and ordinary — knowledge checks in every session and an inspectable-evidence standard. No certificates, no grades, no faith-based claims.
01 · Knowledge Checks
Every session, every time.
Assessment at Primwright is continuous and ordinary — part of every session, not a separate event. Four facts about how it works:
120 checks in Book 1
Twenty-four sessions, five questions each. Every session ends with a knowledge check before the next one begins — comprehension is verified continuously, not at the end.
Judgment, not recall
The checks test practical judgment — classify, verify, draft-and-review, decide — the same operations the session taught. Session 1A's check is an oral classification defense, not a vocabulary quiz.
You verify your own work
The book's answer key is the first verifier. Session 1A's printed answers are on p. 211 of the Learner Edition — no waiting, no grading queue, no black box.
Attempt before reveal
The check is honest by construction: you enter your attempt in your own words before any answer is revealed. An answer you haven't attempted teaches nothing.
02 · How a Knowledge Check Works
Four moves, one honest loop.
This is the exact loop built into the free Session 1A preview — the same loop the books and courses use at every level.
1. Attempt
Answer in your own words — written, spoken, or worked — before seeing anything. The attempt is the evidence.
2. Reveal
The answer unlocks only after your attempt is entered. In the free Session 1A preview, the reveal control stays disabled until you commit to an answer.
3. Verify
Compare your attempt against the book's answer key. Where you diverged, the key shows you exactly where and why.
4. Document
Note what you missed and why. A gap you write down is a gap you can close; a gap you hide is a gap you keep.
03 · Across the Six Levels
Harder objects, same standard.
Assessment deepens with the level — but the rule is constant at every level: inspectable evidence, human judgment, no faith-based claims.
- 01 · Foundations — Level 01 · Beginner — A Beginner's Guide to Understanding, Using, and Verifying Artificial Intelligence
- 02 · Applied AI Systems — Level 02 · Intermediate — An Intermediate Guide to Building, Connecting, Evaluating, and Securing AI Applications
- 03 · AI Engineering & Production Systems — Level 03 · Advanced — An Advanced Guide to Engineering, Securing, Deploying, Observing, and Optimizing AI Systems
- 04 · AI Systems Architecture & Leadership — Level 04 · Architect — An Expert Guide to Designing, Governing, Evolving, and Leading Enterprise AI Systems
- 05 · AI Model Research & Engineering — Level 05 · Research — An Expert Guide to Understanding, Adapting, Evaluating, and Advancing Modern AI Models
- 06 · Frontier AI Systems & Technical Leadership — Level 06 · Frontier — An Expert Guide to Evaluating Frontier Models, Designing Novel AI Systems, Running Rigorous Experiments, Managing Technical Risk, and Leading Technical Research-to-Production Decisions
04 · The Inspectable-Evidence Standard
What we will never do.
The standard stated as commitments — five of them:
- Everything a learner claims is backed by work they can show — attempts and checks.
- Wrong answers are documented, not hidden. The evidence includes the misses.
- No faith-based claims about learning outcomes — no one is told they "know" something they haven't demonstrated.
- The books' answer keys are the first verifier; courses add guided review when they arrive.
- Assessment deepens with the level, but the standard never changes: inspectable evidence, human judgment.
Try It Yourself
From Book 1 · Level 01 · Week 1
Take a real knowledge check.
Session 1A is free in your browser — the session text through the five-question knowledge check. Attempt each question in your own words first; the printed answer key lives in the licensed book (p. 211). That is the whole philosophy, in one session.