Multilingual transcription pilot
Transcription and quality review across multiple locales for a global customer-experience company.
Company
Named for a river that connects.
We connect the physical world, and the people who know it best, to the models that need to learn from it.
Selected work
Recent projects.
Web-scale dataset acquisition pipeline
Designed a cloud pipeline to acquire, filter and store large open video and image corpora with tens of millions of clips and images.
Automated image sourcing at 500K scale
Built a retrieval system that combines a multimodal model with web search to locate and verify source images for a dataset of roughly half a million images.
Study participant operations
Participant tracking and incentive payouts for data collection studies, so research teams can focus on the study itself.
Team
Leaders from Google, Microsoft and TikTok.
Ray Liu
Co-founder17 years at Google, where he led global Trust and Safety for Google Play and Android with 2,000+ vendors, then ran global moderation at TikTok Shop. UC Berkeley EECS.
Dennis Chi
Co-founder18+ years building large-scale systems at Google and Microsoft. Led backend engineering teams for Gmail, Google Workspace and Google Photos. UC Berkeley EECS.