AI Trainer / Data Annotator
As an AI Trainer and Data Annotator at Outlier AI, I developed, evaluated, and refined high-quality prompts and responses used for Large Language Model (LLM) training. My responsibilities included ranking AI-generated outputs, identifying weaknesses like hallucinations or bias, and conducting fact-checking for optimal data accuracy. I collaborated closely with evolving project guidelines to deliver consistent, reliable evaluations and model improvements. • Created and ranked LLM prompts and generated responses for model training. • Identified and documented model errors, biases, and hallucinations. • Applied detailed annotation guidelines and performed adversarial prompting and safety checks. • Conducted rigorous fact-checking and research on model outputs.