Case Study
Training TTS, Audio, and Video models for an award-winning video game
Anuttacon is a leading AI Lab based in San Francisco,and needed subjct matter experts to train their text-to-speech, audio, and video models for their award-winning video game, Whispers From the Star.
The challenge
Anuttacon's models rely on nuanced data that requires human judgement to identify humor, tone, originality, and 'humanness'. Generic annotators or crowdsourced labor were not sufficient, and the client specicially needed subject matter experts from the USA, who understood the culture. They needed experts who could take multi-turn conversations and use their judgement as PhD's in Writing, Linguistics, and Psychology to annotate with the precision required for production-grade models.
Our Solution
Recruitcha supplied subject matter experts from our network, matched to Anuttacon's TTS, audio, and video workflows, as well as from our other networks. We reviewed thousands of specialists, returning a select few after an extensive multi-stage vetting process.
The result
With access to our expert network, Anuttacon scaled their data creation to meet their production deadlines, and delivered a high-quality dataset that helped train models capable of the nuance and quality they needed for their award-winning video game.