AI Training Contributor – Engineering, Data Science & STEM Domains
As an AI Training Contributor at Outlier AI, I created and evaluated datasets to train and benchmark large language models across STEM, engineering, and data science fields. My work emphasized prompt design, rubric validation, response ranking, and chain-of-thought or multimodal reasoning for data quality and alignment. I developed and applied structured annotation frameworks to ensure dataset rigor across real-world technical domains. • Authored multi-step research prompts with verified Golden Truth Final Answers for technical, engineering, and data science subjects. • Performed rubric verification, response ranking, and quality evaluation to enhance AI dataset integrity in STEM domains. • Authored multimodal reasoning questions, including tasks requiring both images and technical solution derivations. • Automated programmatic data generation pipelines for aviation safety, cybersecurity, clinical trials, and real estate compliance.