At Scale AI, I designed and evaluated 500+ AI training tasks, developed grading rubrics, and assessed responses for accuracy, reasoning, relevance, and instruction following. My structured task design and expert feedback contributed to a 70% improvement in AI reasoning evaluation accuracy.
At LXT, I evaluated chatbot responses and reviewed annotated training data to support model improvement. Across my AI training and evaluation work, I investigated hallucinations and other model failures, and quality-assurance processes I implemented reduced annotation and evaluation errors by 85%.
