transcription
As a contributor to Appen’s multilingual text generation project, I worked on creating and reviewing AI training data in both Swahili and Spanish. The scope of the project involved generating prompt-based responses and validating machine-generated outputs to ensure naturalness, fluency, and cultural relevance. The project size was extensive, involving thousands of language-specific prompts designed to enhance the accuracy and contextual understanding of large language models (LLMs). To maintain high-quality standards, we adhered to strict guidelines, including character limits, tone consistency, grammatical accuracy, and contextual appropriateness. Each submission underwent multiple rounds of peer review and quality checks by Appen’s linguistic team. My attention to detail and ability to switch between languages while maintaining linguistic integrity contributed to the overall quality and effectiveness of the dataset.