Assessment
AI/ML skills assessment by conversation, not multiple choice
You can recognise a correct answer about retrieval evaluation without being able to explain it. One of those two gets tested in interviews.
15 free minutes every week
- About 15 minutes
- Voice-first
- No scores
Coverage
3 of 6 areas checked
- Prompt groundingChecked in this session
- Re-rankingChecked in this session
- Failure triageChecked in this session
- Chunking strategyNot yet tested
- Embedding choiceNot yet tested
- Eval harness designNot yet tested
Recognition is not competence, and multiple choice only measures recognition.
Given four options, you will often pick the right one from a half-formed memory of having read something. That feeling registers as knowing. It collapses the moment someone asks you to explain the choice unprompted — which is what an interview does, and what production debugging does. Assessment by explanation is harder to build and considerably harder to fool.
Fifteen minutes later, you have three things.
- A map of what was actually checked
- A gap traced back to what you said
- One thing to study tonight
Explanation as the measurement
You are asked to explain a concept in your own words. There are no options to select between, nothing to eliminate, and no way to pattern-match to a familiar-looking answer. If the explanation thins out under a follow-up, that is the signal.
- No answer options to recognise or eliminate
- Follow-ups probe the specific claim you just made
- Depth is assessed categorically, never as a percentage
Honest coverage reporting
The result separates three states that a single score would blur together: concepts you explained clearly, concepts where the explanation ran out, and concepts the session never reached. The third category is reported as untested rather than being folded into an aggregate that would misrepresent it.
Scoped to AI/ML engineering roles
The assessment covers ML Engineer, LLM Engineer, Applied AI Engineer, AI Infrastructure Engineer, and MLOps roles. It does not assess general software engineering, and it will say so rather than producing a result outside what it can support.
Nothing to install. Talk it through, keep the map.
Before you start
- Is this a test with a pass mark?
- No. There is no pass mark, score, or ranking against other candidates. The output is a coverage map, one possible gap with its evidence, and one next action.
- How is this different from a multiple-choice assessment?
- Multiple choice measures whether you recognise a correct answer. This requires you to produce the explanation yourself and then defend it under a follow-up, which is what interviews and production incidents actually demand.
- Which skills are in scope?
- Those belonging to ML Engineer, LLM Engineer, Applied AI Engineer, AI Infrastructure Engineer, and MLOps work. It is not a general engineering skills assessment, and for a role it has no questions for it returns no result rather than a weak one.
- Is the assessment free?
- Accounts receive 15 minutes weekly on a renewing cycle, no card involved, which covers one assessment start to finish. Assessing more often than weekly costs $29 a month.
Find the concept your prep has not tested.
15 free minutes every week. No card. One gap map, one next action.