AI Benchmark Engineer | Native Language Specialist - German (Switzerland) - Remote

Role overview This freelance opportunity invites a native German-speaking software engineer to help construct a verifiable benchmark suite that probes large language models on multilingual software challenges. The work centers on Terminal-Bench style tasks that evaluate prompt-language effects, non-English data handling, and locale or encoding edge cases in shell-driven workflows. The position is fully remote and structured as independent contract work rather than full-time employment.

Responsibilities - Engineer tasks that probe how coding agents cope with multilingual terminal workflows, avoiding any dependence on English translations. - Build realistic task environments using German-language datasets, file assets, and supporting materials that remain in the target language throughout. - Probe model behavior in German to surface failure points tied to phrasing, register, or grammar rather than to direct translation. - Develop reference implementations and write deterministic verifier scripts, reserving rubric-based judging for cases where strict checks are not feasible. - Run calibration passes across model tiers, adjusting difficulty from easy through very hard using standard Terminal-Bench configurations. - Take part in a four-layer human review process paired with automated checks to safeguard fairness, grammatical accuracy, and benchmark integrity.

Requirements - At least one year of industry experience in software engineering or prompt engineering. - Track record at recognized technology companies and/or a degree from a leading engineering university. - Native or near-native German fluency with strong English proficiency and a deep grasp of German grammar and usage. - Strong command of Python, standard shell scripting, and data processing tooling. - Substantial experience with terminal and CLI-driven development, including working familiarity with AI coding agents. - Deep technical knowledge of multilingual text-processing pitfalls such as Unicode normalization, locale-dependent conventions, safe string handling, and toolchain interoperability.

Benefits and work setup - Fully remote independent contract arrangement with flexible hours and no fixed schedule. - Engagement structured as a 1099 contractor relationship, meaning contractors handle their own taxes and benefits. - Competitive compensation with prompt payment cycles. - Better suited as supplemental income than as a primary job, since project availability varies. - Eligibility is subject to geographic restrictions tied to international sanctions or embargoes.