Constitutional AI SME / RLHF Trainer / Red Team Specialist
As a Constitutional AI SME and RLHF Trainer, I deliver intensive daily red teaming and reinforcement learning from human feedback (RLHF) training on large language models. My work involves identifying and addressing sycophancy, hallucination, bias, and alignment issues to ensure more truthful and empathetic AI systems. I pressure-test AI models while optimizing their responses for factuality, humility, and ethics. • Conducted daily red teaming and RLHF on LLM outputs for safety and alignment. • Optimized for factual, empathetic AI responses by refining training prompts and feedback. • Designed and led technical AI ethics training for engineering and safety teams. • Supported federal clearance projects for sensitive AI safety and alignment initiatives.