The Presented-Work Probe Decoder, Question by Question

You built the deck. You rehearsed the walkthrough. Then the presentation ends and the questions start, and that is where the round is actually decided. This is the lookup table for the recurring follow-up probes that land after a work-sample presentation: one entry per probe family, what each one tests, the trap inside it, and what your answer has to prove instead of describe. It is for anyone about to present a portfolio, a case study, or a shipped project in an interview loop, and it is built to be read mid-prep - jump to the probe you fear and leave with an answer.

Most published advice stops at building and delivering the deck. Almost none of it catalogs the interrogation that follows. That gap matters, because the deck buys you about thirty seconds of credibility and no more.

Why the round inverts once the deck ends

The polished deck helps you for seconds, then works against you. Once the walkthrough is over, evaluation shifts from delivery to interrogation, and rehearsed fluency without decision depth is exactly what the follow-up questions are designed to expose.

A polished portfolio can make almost any candidate look compelling for thirty seconds. The real evaluation starts when the interviewer asks the right follow-up question. That is not a hostile framing - it is the mechanism of the round. The team's job in the Q&A is to push and prod, because a presentation tells them what you built and only the probes tell them how you think.

This inversion has a practical consequence for prep. Time spent making the deck prettier past the point of clarity is time not spent building the decision inventory the probes attack. The deck should be clean enough to not distract. The answers behind it are where the round is won or lost.

The deck buys you thirty seconds of credibility. Everything after that is bought one answer at a time.

The six probe families and what each one scores

Across documented interview sources, the same six families of follow-up recur, and each scores a specific competency rather than testing your recall of the project. Know which family a question belongs to and you know what the answer must prove.

• Rationale probes ("Why did you choose X?", "Why didn't you do Y?") test whether you understood your own decisions or got lucky. You should be able to explain the why behind every design decision and anticipate constraint questions.
• Trade-off probes ("What did that cost you?", "What would break at scale?") test judgment. Two candidates may propose identical architectures, but the one who understands the consequences of the decision demonstrates greater depth.
• Contribution-attribution probes ("What was your specific contribution?") test whether the interviewer can isolate what was actually yours from what the team did.
• Feedback-handling probes ("Describe a situation where you disagreed with feedback you received") test coachability and self-awareness.
• Failure and reflection probes ("Tell me about a time one of your products failed", "What would you change?") test honesty and whether your behavior actually changed.
• Scale and next-steps probes ("What is one improvement you would implement in your first three months?") test forward judgment and whether you see beyond the artifact.

The interviewer is usually less interested in the content of your answer than in what the answer reveals. Feedback probes assess whether you can turn uncomfortable feedback into better performance. Trade-off discussions reveal engineering judgment. Contribution probes exist so the interviewer can understand your specific role and how you made a difference. Creative reviews are rarely about taste; they are about judgment, process, adaptability, and whether you can translate goals into work that performs.

The probe openers and scored competency, by role family

The follow-up sets differ by role family in emphasis but share a common core: rationale, contribution, and reflection appear everywhere, and every family opens with some version of "walk me through a project." Below is the documented opener and the competency it feeds into.

PM loops add strategy and metric probes on top of the core: "What is one improvement you would implement for our product in your first three months?" and "Our metrics are down. How would you determine the root cause?" AI PM interviews now go further and probe evaluation harnesses, offline test sets, online metrics, and model fallback behavior. Design and UX loops lean hardest on constraint and research-method rationale. Whatever the family, the opener is the invitation and the follow-ups are the exam.

Seniority raises the ambiguity bar on the identical question. A "walk me through" answer scoped to a single feature can pass at a mid level and fail at a senior one, where you are expected to run a full product area and mentor others. When you prep, calibrate the scope of your answer to the band you are targeting, not the band you last held.

The decision procedure, step by step

The work that survives the probes is done before the round: you pre-write the why behind every decision, attach a cost to every claim, and rehearse until you can answer without reading. Follow this sequence.

Sources disagree on order in one place. Some put project selection before format alignment; the storytelling-first school treats the narrative of you as the spine and project choice as downstream of it. Either works. What does not work is skipping the trade-off inventory, which is where most weak answers are born.

The answer structure that gets rewarded

The rewarded answer is problem-first, not solution-first: state the constraint that forced the choice, name what the choice gave up, and connect it to the role. Solution-first answers describe; problem-first answers argue.

Compare the two shapes on the same fact. Solution-first: "We used a document database." Problem-first: "We needed to store user profiles with completely different data structures, and a relational schema would have meant constant migrations, which is why we chose a flexible document structure." The second answer proves you understood the problem. The first proves only that you remember the tech.

For "why did you choose this," the three-part frame is: the reason you chose it, what you gained from it, and how it connects to the role. For trade-off probes, add the fourth move that most candidates skip - name the cost. Acknowledging what the choice gives up is the part that signals real understanding, because every real system is a compromise. When you introduce a decision, state the rationale, discuss the limitations, then list the alternatives you considered and why you rejected them.

1. Problem: "The constraint we were under was ___, and if we ignored it, ___ would break."
2. Choice + gain: "So I chose ___, which gave us ___."
3. Cost: "That gave up ___. I accepted that because ___."
4. Alternative rejected: "I also considered ___, but ruled it out because ___."

Several named frameworks package this. CIRCLES structures product-sense answers and includes an explicit "evaluate trade-offs" step. STAR - Situation, Task, Action, Result - is the recurring structure for behavioral and criticism probes. LEARN - Listen, Evaluate, Acknowledge, Respond, Next steps - is a named feedback-handling structure. Use them as scaffolding, not as scripts to recite; an obviously templated answer reads as rehearsed.

The traps: where the honest-sounding answer damages you

Some probes are built so the intuitive, honest-sounding answer is the one that hurts you. These are worth memorizing because you cannot reason your way out of them live.

The feedback probe is the sharpest. Claiming you welcome all criticism without hesitation reads as rehearsed or dishonest, because everyone has an initial emotional reaction to negative feedback. The credible answer is a brief acknowledgment that critical feedback can feel uncomfortable, followed by how you manage that reaction. Two intuitive answers actively damage you: "I just ignore it" tells them you are not coachable, and leading with "I push back if I disagree" makes you sound combative.

The contribution probe has the opposite trap. Generous team credit feels like humility but leaves the interviewer unable to name a single decision that was yours, which fails the probe. Frame the story around you as the protagonist and the specific problem you solved, rather than as a team exercise. Never lie about your role - but how you frame it comes across very differently. The most diagnostic version is a counterfactual: "Would it have been as impactful without you, and why?" That question forces you to isolate your individual causal contribution, which a "we" narrative hides.

Failure modes and false positives

Most people fail these probes in predictable ways, and each failure has a false positive that feels like a good answer while it is sinking you. Here is how to catch each one before the round and correct it live.

• Rationale probe answered as description. The false positive is a fluent walkthrough of what you built. Check: did the answer start with the problem and name a rejected alternative? If not, you are introducing technologies without discussing alternatives.
• Trade-off probe answered with jargon. The false positive is dropping "eventual consistency" or "microservices" without impact. Interviewers probe quickly and shallow explanations are exposed. Check: did you explain what the concept cost you, not just name it?
• Contribution probe answered in "we." The false positive is generous team credit that reads as humility. Check: can the interviewer name one decision that was yours? If not, switch to "I" for your own choices.
• Feedback probe over-polished. The false positive is "criticism doesn't bother me." It sounds dishonest. Check: did you acknowledge an initial reaction before the behavior change?
• Feedback probe with a trivial example. The false positive is a safe, tiny critique. A very small critique feels like avoiding the question and the hiring manager assumes you are hiding something. Check: did the feedback have real stakes?
• Failure probe with a disguised strength. Check: is it a real, meaningful weakness that had consequences, not a disguised strength or a trivial nit?
• Feedback probe that blames the giver. The false positive is "it wasn't fair criticism." Check: did you avoid talking poorly about the person, even if the delivery was unfair?
• AI or technical probe with a demo but no evaluation. Check: did you describe how you measured it? A model that demos beautifully with no evaluation behind it is a liability with a launch date.

The tells cluster by family. This table maps each probe family to its documented weak-answer signature so you can hear yourself sliding into one.

The single most useful diagnostic is grammatical. Padders name the product; builders name the decision. If your answer to "what was your contribution" is the name of the feature, you are describing. If it is the specific decision you made that caused the feature to succeed, you are proving ownership.

Where the bench is deep, and how to rehearse against it

The depth of the follow-up bench you face scales with how many people in the field can interrogate your work, and that varies by market. In Refolk's index of professional profiles there are 16,369 US Product Designer or UX Designer profiles versus 4,053 in the UK - roughly a 4.0x supply gap - so a US design candidate is more likely to be probed by an experienced practitioner who has seen your genre of project before.

Note the limits of this number. It measures supply of people with those titles in Refolk's index, not interviewer quality or how hard any given loop probes. I queried senior-band figures for Software Engineer and Product Manager in the same index and the title-plus-seniority filter returned no matches, so I will not claim a seniority-based supply comparison here. Treat the country comparison as a directional read on bench depth, not a precise forecast of your loop.

The practical move is to rehearse against a real practitioner in your role family before the round, so the traps get sprung in practice rather than live. A mock reviewer who has run portfolio critiques will fire the counterfactual contribution probe and the "name the cost" trade-off probe that a friend will not think to ask. Refolk can find that person by description rather than by you already knowing who they are.

If you would rather practice the written spine of the answer first, Refolk can draft the problem-first framing for each project from your own history, so you walk in with a rationale and a rejected alternative already attached to every visible decision. That converts the prep in step three from a blank page into an edit.

Before you walk in

Run this checklist against every project you plan to present. If any item fails, you have a probe you cannot yet answer.

The probes do not change much year to year. What changes is whether you walked in with a cost attached to every claim and a decision you can call your own. Build that once, keep it in a document you update after every loop, and the interrogation stops being the part you fear.