AI Trainer (Remote Freelance) | Deccan AI
As an AI Trainer at Deccan AI, I contributed extensively to the training and evaluation of large language models (LLMs). My responsibilities included data annotation, prompt engineering, evaluation of AI responses and agentic actions, and reinforcement learning from human feedback (RLHF). This work enhanced AI safety, multi-step reasoning, and the overall performance of deployed agents. • Executed high-quality data labeling and annotation tasks on textual datasets for machine learning. • Designed and refined prompts and evaluated model completions for appropriateness and correctness. • Performed human-in-the-loop RLHF annotation cycles to improve model reliability and alignment. • Collaborated in continuous evaluation of agentic actions, optimizing AI tool usage and safety protocols.