Applied Scientist - Deep Learning

Role overview This is an applied scientist role focused on deep learning for audio processing within an early-stage voice AI startup. The scientist will build state-of-the-art audio models, fine-tune foundation models, develop benchmarks, construct data pipelines, and run experiments to drive continual improvement. The work targets reliable voice AI in live restaurant environments such as drive-thrus.

Responsibilities - Build state-of-the-art audio processing deep learning models at small and larger scales - Fine-tune and post-train foundation models - Build benchmarks that align with real customer needs - Construct and maintain data pipelines - Run experiments to continually improve model and system performance

Requirements - PhD in a technical field or equivalent applied deep learning experience - Hands-on experience training deep learning models

Nice to have - Experience working with PyTorch - Experience building neural audio processing models

Benefits and work setup - Fully remote within the United States - Visa sponsorship is not offered - Remote-first startup with a distributed virtual office culture - Open to candidates across the U.S. who are authorized to work there