Appen
Appen collects, annotates and curates training data for machine learning through a contributor network spanning 130 countries and 180 languages, and also offers model fine-tuning and evaluation benchmarks.
Appen supplies the human labour behind artificial intelligence. Founded in 1996 and tracing its origins to a linguistics project, the company collects, annotates and curates the data on which machine learning models depend, selling its services to some of the world's largest technology firms. Its work spans speech recognition, computer vision, large language models and generative AI.
The company's main service is the human-driven collection, annotation and curation of training data, intended to make models function under real-world conditions. It is delivered through a distributed contributor network that reaches 130 countries and 180 languages. Two further lines of work are model fine-tuning for clients, with an emphasis on precision and cultural relevance, and the creation of new evaluation benchmarks designed to test whether AI systems behave reliably across populations and settings.
Appen has been in operation for more than 25 years. Its stated priorities are quality and accuracy over sheer data volume, cultural relevance in the data it produces, and building AI that performs consistently for users everywhere. The workforce model is global and distributed rather than office-based, reflecting the geographic spread of its contributor network.