Reading a task rubric without losing points
· a.pannella87
Last updated: 27 October 2026 · OpenJob editorial team
To avoid losing points, read a task rubric twice: first to list the decisions you make on every item, then to find the exceptions marked by words like unless, except and even if. Keep a one-page cheat sheet of rules you tend to forget, and run the project’s quality checklist on your own work before submitting.
Key points
- First read: the decisions you will be scored on.
- Second read: the exceptions, where reviewers catch people.
- Date your cheat sheet; guides change.
- When the guide is silent, follow the closest example and say so.
Every AI data project comes with a guide, and most people read it once. The workers whose tasks pass review read it differently: as a reference they return to, not a briefing they finish.
First read: find the decisions
Skim the whole guide and list the decisions you will make on every item, for example “is the answer accurate?”, “which category?”, “is it safe?”. Those are the headings your work will be scored on.
Second read: find the exceptions
Look for words like “unless”, “except” and “even if”. Exceptions are where reviewers catch people, because they contradict intuition.
Keep a personal cheat sheet
Write the five rules you are most likely to forget on one page. Update it whenever a reviewer corrects you. Projects change their guides; date your sheet so you know when to recheck it.
When the guide is silent
- Look for the closest example in the guide and follow its logic.
- If none fits, choose the more conservative option and say so in your comment.
- Ask in the project’s question channel, quoting the item, so the answer helps everyone.
Quality checklists
Many projects publish a checklist reviewers use. If you get one, run it on your own work before submitting. It is the closest thing to seeing the marking scheme.
This habit pays off on every kind of listing, from usability testing to specialist evaluation. Browse open roles.
Open roles that do not ask for previous experience
Refreshed automatically from our verified catalogue. Check eligible countries and requirements on each listing.
Corporate Attorney (BigLaw Firms)
Remote contract work for corporate attorneys from large law firms on an AI training project: analyze and draft complex agreements and legal instruments in corporate law, M&A, private equity and cap…
Finance Specialist
Remote contract for finance professionals on an executive-presentation project: rebuild and improve PowerPoint slides that present complex financial data and insights for investors and executives…
Bilingual Japanese Psychologist (PhD)
Remote contract for bilingual psychologists: assess psychological case studies and documentation and develop culturally sensitive scenarios and materials in Japanese and English for AI training.
Simulation Expert — Real-to-Sim 3D Environments & Assets
Remote contract work for simulation experts on an AI training project: build 3D assets and environments in Isaac Sim, Gazebo, MuJoCo or similar, and tune physical parameters so they match real-worl…
Japanese Language Expert
What if your mastery of Japanese could directly influence how AI understands, speaks, and connects with millions of people?
QA and AI evaluation roles · All open roles
Frequently asked questions
What should I do when the guidelines do not cover a case?
Find the closest example in the guide and follow its logic. If none fits, choose the more conservative option and explain it in your comment. Then ask in the project’s question channel, quoting the item, so the answer helps everyone working on it.
How often do task guidelines change?
Often, especially on long projects. Clients refine rules as they see results. Date your personal cheat sheet, reread the changelog or announcements when they appear, and recheck rules after any reviewer correction. Working from an old version is a common cause of lost points.
Is a quality checklist the same as the rubric?
Not exactly. The rubric defines how items are judged; a checklist is the list reviewers use to check your work. If a project shares one, run it on your own tasks before submitting. It is the closest thing to seeing the marking scheme.