
AI Benchmark Engineer, Native Language Specialist – Portuguese (Brazil)
Posted Sep 9

Posted Sep 9
This is a fully remote position, open to applicants in Brazil.
• Design, construct, and validate Terminal-Bench tasks that assess coding agents.
• Create realistic task environments utilizing datasets and files in Portuguese, ensuring that assets remain in the target language.
• Identify AI failure points through prompting and translation in Portuguese.
• Develop strong reference implementations.
• Write dependable, deterministic verifier scripts, employing rubric-based evaluation only when absolutely necessary.
• Examine execution logs and adjust task difficulty from Easy to Very Hard.
• Execute standard Terminal-Bench configurations against Haiku, Sonnet, and Opus model tiers.
• Engage in a four-layer human quality-control process that includes creation, human review, calibration review, and audit.
• Collaborate with automated LLM-based checks to guarantee fairness, grammatical correctness, and benchmark integrity.
• A minimum of 1 year of industry experience in software or prompt engineering.
• Demonstrated success at leading technology firms and/or graduation from prestigious engineering universities.
• Native or near-native fluency in Portuguese (Brazil) with a profound understanding of grammar, register, and phrasing rules.
• High proficiency in English.
• Strong expertise in Python, standard shell scripting, and data processing.
• Significant experience with Terminal/CLI-based development workflows.
• Familiarity with coding agents.
• In-depth technical knowledge of multilingual text processing challenges.
• Experience with encoding/decoding robustness and Unicode normalization.
• Understanding of locale-dependent conventions, including collation, casing, and non-Gregorian dates.
• Knowledge of text I/O, toolchain interoperability, and safe string operations.
• An updated CV in English.
• Completion of a GenAI assessment is required.
• Reliable availability and commitment; most tasks necessitate at least 2 hours per day or 10 hours per week.
• Contractors must not be engaged in regions subject to international embargoes or sanctions.
• Contractors are accountable for their own tax obligations.
• Flexible schedule without fixed hours, check-ins, or micromanagement.
• Competitive rates with prompt payments.
• Access to a variety of innovative projects that enhance your portfolio and develop your skills.
• A global community of linguists, subject matter experts, and language professionals.
• No benefits such as health insurance, paid time off, or retirement contributions are provided.
• Work availability may vary based on project demand, and hours are not guaranteed.
Mercor
Mercor
Triple Whale 🐳
Get handpicked remote jobs straight to your inbox weekly.