AI Trainer Intern – Video Data Labeling and Prompt Optimization
During my internship as an AI Trainer, I was responsible for designing and refining prompt instructions, optimizing data production workflows, and ensuring quality outputs for video generation projects. I produced and labeled AI training data, including single/multi-character and multi-camera randomized walkthrough video sequences, passing rigorous quality assurance checks. I leveraged Python automation for batch file handling, efficiency enhancements, and standardized parameter configurations. • Wrote and debugged prompt instructions for Unreal Engine 5 scene rendering for diverse AI training data production. • Generated and labeled videos and corresponding JSON label files in batches that passed 100% of quality inspections. • Developed automation scripts with Python to streamline folder creation and parameter standardization, increasing efficiency by over 60%. • Built and optimized prompt templates using Coze intelligent agents for parameter validation and workflow tracking.