About the Role
As an AI Trainer specializing in LLM annotation and evaluation, you will be responsible for annotating, reviewing, and scoring large language model outputs focused on STEM reasoning, coding logic, and multiple-choice question evaluation. You will ensure high-quality training data and accurate model assessment, directly supporting the improvement of AI agents technical reasoning and coding capabilities.
About You
Education & Background
- Diploma or above in a STEM-related field
- Preferred majors: Computer Science, Software Engineering, Information Technology, Electronic Engineering, Mathematics, Statistics, Data Science, Artificial Intelligence, or equivalent professional certifications.
Core Competencies
- Basic proficiency in Python syntax: able to independently read and understand pseudocode and Python code snippets, trace execution results, judge logical correctness, and accurately identify correct answers for MCQs.
- English literacy at CEFR B1 level or above: able to fully understand English instructions, annotation guidelines, project requirements, and question prompts without comprehension errors.
- Able to perform annotation and evaluation on coding, logical reasoning, and STEM-related tasks with precision and consistency.
Preferred Qualifications
- Previous experience in Python programming, programming education, question bank verification, or AI training data annotation (especially coding tasks or MCQ evaluation).
- Familiarity with LLM-related work such as RLHF, model evaluation, data auditing, or annotation projects.
- Basic hands-on coding experience and strong logical thinking ability.
Responsibilities
- Annotate and evaluate model performance on coding, logic, reasoning, and STEM-related tasks.
- Review Python code snippets, trace logic, judge correctness, and complete MCQ evaluation accurately.
- Construct and organize high-quality evaluation datasets for STEM, coding, and reasoning scenarios.
- Follow English project guidelines strictly to ensure annotation quality and data consistency.
- Complete daily, weekly, and monthly annotation/evaluation tasks as assigned.
