A.I Trainer

Responsibilities:

• Architect and manage end-to-end training workflows, including supervised fine-tuning (SFT), Reinforcement Learning from Human Feedback (RLHF), and red-teaming exercises.
• Develop complex prompt templates and testing scenarios to evaluate model performance across various domains (coding, creative writing, reasoning, and multilingual tasks).
• Define and maintain rubrics for model output quality. Conduct in-depth analysis of model failures and edge cases to identify root causes of bias, hallucination, or logical errors.
• Create high-quality “golden datasets” that set the standard for model behavior in your area of expertise.
• Guide and review the work of junior trainers and data annotators, ensuring consistency and high standards across the global team.
• Maintain detailed documentation of training methodologies, model behaviors, and iterative improvements for internal stakeholders and research teams.
• Work closely with Machine Learning Engineers and Product Managers to translate technical research goals into actionable training tasks for human annotators.
• Identify barriers in the training workflow and propose improvements to the internal platforms or tools used for data annotation and model evaluation.
• Design and implement strategies for generating high-quality synthetic data to augment training datasets and improve model robustness in niche scenarios.

Requirements:

• Minimum of 4 years of experience in a relevant field (e.g., Software Engineering, Computer Science).
• Proven ability to dissect complex problems, identify patterns in data, and articulate abstract concepts clearly.
• Familiarity with Python, JSON, or APIs is a plus. A willingness to learn internal tooling and understand basic machine learning concepts is required.
• Comfortable reading technical papers or documentation to understand how models are expected to behave and staying updated on the latest A.I. safety research.
• Experience using data analysis or metrics to justify training decisions and measure the success of specific interventions on model performance.
• Ability to manage multiple concurrent projects, prioritize tasks effectively in a fast-paced environment, and meet tight deadlines.
• Willingness to shift focus between highly technical tasks (e.g., ranking code) and creative tasks (e.g., writing fiction) as project needs evolve.

Remuneration: NGN 500,000 monthly