AI Trainer / Domain Expert (Computer Science & Software Engineering)
Text
RLHFAs an AI Trainer and Domain Expert, I authored and reviewed thousands of high-quality prompt-and-response pairs for large language models. I graded model outputs against multi-dimensional rubrics focusing on correctness, reasoning, code quality, and safety, producing detailed critiques. I conducted adversarial prompting, red-teaming, and created demonstrations for optimal model behavior. • Developed supervised fine-tuning data and preference data for software, algorithms, system design, and ML domains. • Created and applied rubrics to systematically evaluate LLM outputs. • Identified model failure modes through structured adversarial testing. • Reviewed and mentored peers to ensure high-quality data and onboarding of new trainers.
Present