The vocabulary the exams and the daily work both assume you already have. Every entry gives the definition, the source it is drawn from, and the misreadings that follow from getting it wrong.
265 concepts, page 7 of 14
Compare JSON prompting, JSON mode, schema constrained output and function calling. Learn what each guarantees and how to handle incomplete responses.
Artificial Intelligence
Understand temperature, top p, top k and seeds in LLM generation, including why temperature zero does not guarantee reproducible responses.
Artificial Intelligence
Use an LLM judge to score open ended responses with a clear rubric, then measure agreement, false passes and position bias against human labels.
Artificial Intelligence
Disagreement about the work is information the project needed, disagreement about people is worth stopping, and the five conflict modes exist because no single response fits every case.
Project Management
A problem held by many people in a defined market, stated without a solution attached, and qualified on urgency, pervasiveness and willingness to pay.
Discovery
Compare customer segments by needs, buying behaviour, reach and economics, then choose a practical first market to test.
Strategy
Decisions made from evidence gathered outside the building rather than from opinion gathered inside it, and the specific habits that produce that evidence.
Strategy
Acceptance rate and the unit it is counted in, edit and override rate, the error rate a team has to build because nothing emits an event for a wrong answer, the arithmetic of the trade between latency and quality, and why ordinary engagement measures point the wrong way for a generative feature.
Metrics
Value is established by evidence rather than assertion, and the Increment is what makes it inspectable.
Scrum
What a model retains from its training data, how that can be pulled back out, and what differential privacy does and costs.
Artificial Intelligence
Learn when multiple AI agents help, how to design subtask contracts and how to measure handoff quality, shared budgets and coordination costs.
Artificial Intelligence
Why examining many measures at the five per cent threshold produces winners out of nothing, the arithmetic for twenty measures and for ninety six segment cuts, the two corrections available and the practice that avoids needing either.
Metrics
Scope, resources, schedule and priority inside the project manager's own organisation account for most of what gets negotiated on a project, and almost none of it is called negotiation.
Project Management
The effect that fades once a change stops being new, the effect that grows once people have learned their way around it, the two readings that separate them from noise and the published evidence on how often either one decides anything.
Metrics
The spectrum from a closed hosted model to a fully open release, the arguments on both sides, and what changes when a team builds on weights it can download.
Artificial Intelligence
Comparing candidate opportunities against consistent criteria so the choice is made on the same basis every time, and so a rejected opportunity can be shown why it was rejected.
Prioritisation
Order expresses strategy, not preference, and being accountable for it is not the same as doing the work.
Scrum
Getting people to work differently once the thing is built, which is where most of a project's value actually sits and is a separate discipline from controlling changes to the baseline.
Project Management
Output is what the team produced. Outcome is what changed for someone because of it. A product owner is accountable for the second and is usually measured on the first.
Metrics
Why people accept output they should have checked, why it gets worse as the system gets better, and what a product can do about it.
Artificial Intelligence