Remote Multilingual AI Benchmark Engineer

LILT AI; Pontianak; Remote; IDR 992,686,000 - 1,985,372,000; Part time

Benefits offered by this job

Flexible schedule

Competitive pay

Global community

Remote freelancing

Job summary

LILT recruits experienced native-speaking software engineers to design, build, and validate benchmarks for multilingual model evaluation. You will create high-signal tasks and environment assets in your native language to measure robustness without English crutches.

This remote freelance role emphasizes rigorous quality control, deterministic verification scripts, and scalable calibration across model tiers. Flexible schedule and independent contractor setup offered.

Qualifications

• 1+ years of industry software engineering experience.
• Graduation from top-tier tech universities or equivalent.
• Native or near-native fluency in language with high English proficiency.
• Proficient in Python, shell scripting, and data processing.
• Experience with Terminal/CLI workflows and coding agents.
• Knowledge of multilingual text processing pitfalls and encoding issues.

Responsibilities

• Design, build, and validate benchmarks for multilingual models.
• Create task environments and assets in target languages.
• Identify failure points via prompting and translation in native language.
• Develop reference implementations and robust verifier scripts.
• Calibrate task difficulty across model tiers and run configurations.
• Participate in a 4-layer QA process ensuring fairness and quality.

Skills

Python

Shell scripting

Data processing

Multilingual NLP

Terminal/CLI workflows

Coding agents

Education

Graduation from top-tier engineering universities

Job description

LILT recruits experienced native-speaking software engineers to design, build, and validate benchmarks for multilingual model evaluation. You will create high-signal tasks and environment assets in your native language to measure robustness without English crutches.

This remote freelance role emphasizes rigorous quality control, deterministic verification scripts, and scalable calibration across model tiers. Flexible schedule and independent contractor setup offered.