
AI Trainer (Contractor), Voice and Speech Evaluation - USA
Inworld AI Mountain View
About the role
At Inworld AI, a company specializing in real-time AI for consumer applications, the AI Trainer role is part of the operations team focused on voice and speech evaluation. The company develops speech models and APIs supporting diverse sectors like healthcare, education, and media, serving billions of users daily. This contractor position is remote within the USA and involves working with cutting-edge speech technology and evaluation processes.
The AI Trainer's daily tasks include listening to speech samples to assess naturalness, clarity, and emotional expression, scoring voices against detailed rubrics, and labeling voice attributes such as gender and accent. They also review new voice batches for quality, test speech features in pre-release environments, and provide clear, actionable feedback to engineers. Multilingual candidates may evaluate pronunciation and dialect accuracy in their native language.
Candidates should have a background in audio-related fields like linguistics or voice acting, with a keen ear for subtle audio differences and consistent judgment across many samples. Comfort with web tools, VPNs, and spreadsheets is required, along with strong written English skills and the ability to meet tight deadlines. Preferred qualifications include fluency in certain languages and experience with TTS, voice AI, or localization quality assurance.