Join a short, paid project that tests whether AI coding agents work as well on instructions written in the target language as they do in English. Your language skills and technical background decide where you fit.
About the project
Welo Data is helping evaluate how well AI coding agents perform when the instructions they receive are written in languages other than English. Agents are given a real software engineering task in a code repository, and automated tests check whether they solved it. The code, repositories and tests stay exactly as they are. Only the human-written instructions change language, so the quality of every instruction directly affects the result.
About the roles
There are two roles on this project. We place you in one of them based on your experience & skillsets.
Translator: translate English software engineering task briefs into the target language. You keep every requirement, constraint and success condition, and you leave commands, file paths, API names, code and test conditions exactly as written in English. This role suits people with a stronger technical background, and translation experience is a plus.
Reviewer: compare each translation with the English source, check that nothing is missing, added or changed in meaning, correct errors, and log each correction by error type. This role suits people with strong translation, editing or review experience and enough technical familiarity to read code context.
Project at a glance
Project duration: one week since the mid of November
Language: German
Project time commitment: 6 hours per person
Project compensation: US$72.5 - US$87.5 per hour
Work location: fully remote in Germany