AI Trainer (SWE)

About the Role

We are looking for skilled Software Engineers to contribute to a project focused on the advancement of sophisticated AI technologies. This role is intended for experts with a background in AI model evaluation or training who are prepared to utilize their technical proficiency to enhance the efficacy of large-scale models. You will play a vital part in advancing AI capabilities by maintaining high benchmarks for technical complexity, accuracy, and robustness through detailed assessment and testing. Candidates should excel in a setting that prioritizes critical analysis, precision, and extensive familiarity with software development lifecycle methodologies. This role provides a flexible schedule for contractors capable of contributing a minimum of 15 hours weekly. Through your efforts on intricate engineering projects, you will shape the progress of AI instruments designed for complex professional development environments.

Requirements

• You must hold verifiable professional experience in benchmarking, evaluating, or training artificial intelligence systems.
• You are required to be proficient in Python and have professional experience with at least one additional language from the following: JavaScript, Java, TypeScript, C#, C++, Rust, or Go.
• You must possess comprehensive understanding of dependency management, version control, testing, debugging, and software design principles.
• You must submit proof of your technical expertise via resources such as GitLab or GitHub profiles, technical papers, competitive programming achievements, or contributions to open-source projects.
• It is preferred that you have experience with agentic coding evaluations, familiarity with Docker and Linux, a strong record in competitive programming, and background in designing grading tasks or evaluation benchmarks.

Responsibilities

• You will design authentic software engineering problems to challenge the capabilities of cutting-edge AI systems.
• You will formulate detailed instructions, configure reproducible settings, and produce reference solutions complete with necessary tests and grading metrics.
• You will review code generated by AI to determine its completeness, robustness, and accuracy.
• You will diagnose specific errors involving debugging, implementation, testing frameworks, dependencies, and performance to ensure models meet professional quality standards.

Benefits & Perks

• This is a fully remote role featuring a flexible schedule with a mandatory minimum of 15 hours of work per week.
• Compensation is based on project output, offering an hourly equivalent rate between $60 and $120, determined by the quality of your work and the complexity of the assignments.