Travel Expert | Reviewer

<p data-start="0" data-end="160">We’re recruiting <strong data-start="17" data-end="71">Travel professionals and experienced trip planners</strong> to review and improve <strong data-start="94" data-end="124">realistic travel scenarios</strong> used to train AI travel assistants.</p> <p data-start="162" data-end="383">Your job isn’t to “write itineraries.” It’s to think like a real trip planner and spot what makes a travel situation <strong data-start="279" data-end="320">real, feasible, and decision-changing</strong> — then upgrade weak scenarios so they actually test the model.</p> <p data-start="385" data-end="604">‼️ <strong data-start="388" data-end="410">AI is not allowed.</strong> If we spot AI use, we’ll block the application.<br data-start="458" data-end="461">⚠️ <strong data-start="464" data-end="528">This application form must be completed entirely in English.</strong><br data-start="528" data-end="531">🚨 <strong data-start="534" data-end="604">READ ALL THE DESCRIPTION BELOW TO UNDERSTAND THE APPLICATION FORM.</strong></p> <hr data-start="606" data-end="609"> <p data-start="611" data-end="624"><strong>WHO CAN APPLY</strong></p> <ul data-start="625" data-end="812"> <li data-start="625" data-end="677"> <p data-start="627" data-end="677">Travel professionals and experienced trip planners</p> </li> <li data-start="678" data-end="766"> <p data-start="680" data-end="766">3+ years of relevant travel planning experience (professional or high-volume personal)</p> </li> <li data-start="767" data-end="812"> <p data-start="769" data-end="812">English level: Advanced / Fluent (required)</p> </li> </ul> <hr data-start="814" data-end="817"> <p data-start="819" data-end="1011"><strong>WHAT YOU’LL DO IN THE APPLICATION (QUICK SCREENING TEST)</strong><br>You will review <strong data-start="892" data-end="926">two pre-generated travel tasks</strong>. For each task, you’ll answer simple <strong data-start="964" data-end="979">1–5 ratings</strong> and give a short justification.</p> <p data-start="1013" data-end="1083">Then, for the <strong data-start="1027" data-end="1039">2nd task</strong>, you’ll <strong data-start="1048" data-end="1082">rewrite it to make it stronger</strong>.</p> <hr data-start="1085" data-end="1088"> <p data-start="1090" data-end="1299"><strong>WHAT “STRONG” MEANS (SIMPLE DEFINITION)</strong><br>A strong task includes <strong data-start="1153" data-end="1192">one decisive feasibility constraint</strong> that can flip the plan if ignored (missed train, missed check-in, closed attraction, wrong station, etc.).</p> <p data-start="1301" data-end="1326">We call this a <strong data-start="1316" data-end="1325">Knife</strong>:</p> <ul data-start="1327" data-end="1479"> <li data-start="1327" data-end="1404"> <p data-start="1329" data-end="1404">A Knife is a constraint that <strong data-start="1358" data-end="1403">changes what itinerary is actually viable</strong>.</p> </li> <li data-start="1405" data-end="1479"> <p data-start="1407" data-end="1479">Quick test: if you remove it, the “best plan” would meaningfully change.</p> </li> </ul> <p data-start="1481" data-end="1509">Examples of Knives in Travel</p> <ul data-start="1510" data-end="1794"> <li data-start="1510" data-end="1559"> <p data-start="1512" data-end="1559">Last train / last entry / check-in cutoff times</p> </li> <li data-start="1560" data-end="1611"> <p data-start="1562" data-end="1611">Closures and holidays that change what’s possible</p> </li> <li data-start="1612" data-end="1678"> <p data-start="1614" data-end="1678">Ambiguous locations (airport/station names that can be confused)</p> </li> <li data-start="1679" data-end="1739"> <p data-start="1681" data-end="1739">Sequential dependencies (if A slips, B becomes impossible)</p> </li> <li data-start="1740" data-end="1794"> <p data-start="1742" data-end="1794">Mobility constraints that change routing feasibility</p> </li> </ul> <hr data-start="1796" data-end="1799"> <p data-start="1801" data-end="1869"><strong>WHAT YOU WILL RATE (1–5) IN THE FORM</strong><br>You’ll give 3 ratings per task:</p> <ol data-start="1871" data-end="2349"> <li data-start="1871" data-end="1959"> <p data-start="1874" data-end="1959">Overall Quality (1–5)<br data-start="1895" data-end="1898">How clear and testable the task is (can you judge PASS/FAIL)?</p> </li> <li data-start="1961" data-end="2122"> <p data-start="1964" data-end="2122">Knife / Hidden Context Strength (1–5)<br data-start="2001" data-end="2004">Is there a decisive feasibility constraint — or is the “hidden context” just extra color that doesn’t change the plan?</p> </li> <li data-start="2124" data-end="2349"> <p data-start="2127" data-end="2349">Checklist Quality (1–5)<br data-start="2150" data-end="2153">Does the checklist truly test feasibility + the Knife with clear criteria?<br data-start="2227" data-end="2230">(Strong checklists include verification steps and forbid inventing logistics, written as PASS/FAIL — not “be helpful.”)</p> </li> </ol> <p data-start="2351" data-end="2429">You’ll also write a short justification (2–4 bullets) explaining your ratings.</p> <hr data-start="2431" data-end="2434"> <p data-start="2436" data-end="2537"><strong>WHAT YOU WILL REWRITE (FOR THE 2ND TASK ONLY)</strong><br>You will rewrite these parts to make the task stronger:</p> <ul data-start="2538" data-end="3127"> <li data-start="2538" data-end="2621"> <p data-start="2540" data-end="2621">Scenario (3rd person) — more realistic + naturally sets up the feasibility risk</p> </li> <li data-start="2622" data-end="2691"> <p data-start="2624" data-end="2691">Prompt (user message) — natural, human, does not reveal the Knife</p> </li> <li data-start="2692" data-end="2764"> <p data-start="2694" data-end="2764">Hidden Context (1–3 bullets) — must include exactly 1 decisive Knife</p> </li> <li data-start="2765" data-end="3127"> <p data-start="2767" data-end="2829">Checklist (4–6 items) — each item must be PASS/FAIL, covering:</p> <ul data-start="2832" data-end="3127"> <li data-start="2832" data-end="2886"> <p data-start="2834" data-end="2886">feasibility/verification (minimum unknowns to check)</p> </li> <li data-start="2889" data-end="2935"> <p data-start="2891" data-end="2935">the Knife (the decision-flipping constraint)</p> </li> <li data-start="2938" data-end="3003"> <p data-start="2940" data-end="3003">no invented logistics (no made-up schedules/hours/availability)</p> </li> <li data-start="3006" data-end="3087"> <p data-start="3008" data-end="3087">branching plan (Plan A / Plan B) when feasibility depends on one missing answer</p> </li> <li data-start="3090" data-end="3127"> <p data-start="3092" data-end="3127">safety only when there is real risk</p> </li> </ul> </li> </ul> <p data-start="3129" data-end="3290">Checklist item example:<br data-start="3152" data-end="3155">PASS if it flags what must be verified (tickets/hours/cutoffs) before committing; FAIL if it confidently invents times or availability.</p> <p data-start="3292" data-end="3395">Important: Don’t “leak” the Knife in the prompt (no “IMPORTANT”, no all-caps, no “the key detail is…”).</p> <hr data-start="3397" data-end="3400"> <p data-start="3402" data-end="3485"><strong>IF SELECTED (PROJECT WORK)</strong><br>You’ll review batches of pre-generated travel scenarios:</p> <ul data-start="3486" data-end="3713"> <li data-start="3486" data-end="3523"> <p data-start="3488" data-end="3523">Rating realism and feasibility risk</p> </li> <li data-start="3524" data-end="3574"> <p data-start="3526" data-end="3574">Identifying weak or missing constraints (Knives)</p> </li> <li data-start="3575" data-end="3713"> <p data-start="3577" data-end="3713">Improving tasks so they better represent real travel moments where wrong advice causes missed connections, wasted money, or ruined plans</p> </li> </ul> <hr data-start="3715" data-end="3718"> <p data-start="3720" data-end="3873"><strong>COMPENSATION</strong><br></p> <p>Payment will be US$ 30 per approved task, converted and paid in your local currency.</p> <p>Each task takes approximately 40 minutes, which corresponds to an effective rate of about US$ 45 per hour.</p> <hr data-start="3875" data-end="3878"> <p data-start="3880" data-end="3902"><strong>WHAT WE’RE LOOKING FOR</strong></p> <ul data-start="3903" data-end="4167" data-is-last-node="" data-is-only-node=""> <li data-start="3903" data-end="3970"> <p data-start="3905" data-end="3970">You can recognize what’s realistic vs. generic in travel planning</p> </li> <li data-start="3971" data-end="4071"> <p data-start="3973" data-end="4071">You understand what truly flips feasibility (cutoffs, closures, dependencies, ambiguous locations)</p> </li> <li data-start="4072" data-end="4167" data-is-last-node=""> <p data-start="4074" data-end="4167" data-is-last-node="">You can upgrade a “fine” scenario into a strong feasibility test without making it artificial</p> </li> </ul>