Base Salary
$180k - $350k/yr
Responsibilities
- Build software systems supporting distributed model training, inference, and benchmarking.
- Develop systems for massive-scale data collection, storage, preprocessing, and analysis.
- Solve advanced AI research problems at industry scale.
- Contribute to the development of speech-language models optimized for human expression and well-being.
Requirements
- Expertise in the Python ecosystem and popular machine learning libraries and tools, such as PyTorch.
- Experience writing robust, maintainable, production-ready code.
- Ability to iterate quickly on new and uncertain research directions.
- At least two years of experience training and/or fine-tuning transformer models with large-scale text, audio, image, and/or video datasets.
Benefits
- Work location is in New York City.
Categories
About Hume AI
Built from a decade of voice and emotion research. To build emotionally intelligent voice AI, we first had to define what “good” sounds like – not just acoustically, but perceptually and in real human conversations. That led us to build the research infrastructure behind expressive, trustworthy voice AI: scientifically grounded datasets, speech models, evaluation frameworks, and human preference pipelines. Today, we make that infrastructure available to frontier labs and AI-native companies building the next generation of voice. Whether you’re building foundation models, fine-tuning voice agents, or evaluating production systems, Hume provides the tools to measure, improve, and align voice AI the way people actually experience them.