Outlier AI
At Outlier AI, I worked on AI training and evaluation projects where I assessed and compared AI-generated text and image outputs to improve overall model performance and response quality. My work involved prompt evaluation, response ranking, and comparative analysis to determine accuracy, relevance, coherence, and instruction adherence across multimodal tasks. I identified inconsistencies, contextual errors, and low-quality outputs while following structured evaluation guidelines to support the development and optimization of large language models and generative AI systems.