Role Overview
Hume AI is hiring a full-time Research Engineer to advance speech-language and audio models through hands-on development, large-scale training, and testing. This on-site position is located at Hume AI HQ and suits individuals eager to build artificial intelligence optimized for human well-being.
Key Responsibilities
Design and execute experiments to test hypotheses regarding speech-language, audio, and multimodal machine learning model behaviors.
Train, fine-tune, validate, and develop proprietary systems alongside customer-built models.
Construct stable software to facilitate distributed model training, inference, and benchmarking for reproducible results.
Establish data workflows to handle the collection, storage, preprocessing, and analysis of large-scale datasets.
Create evaluation protocols and benchmarks to measure model quality, robustness, and generalizability.
Examine datasets, model outputs, and failure points to guide subsequent development cycles.
Partner with external technical and research teams to define objectives and execute validation and training programs.
Document all experimental designs, outcomes, and decisions for team and customer transparency.
Requirements
Minimum of two years of practical experience contributing to the training or fine-tuning of large-scale text, audio, image, or video datasets, with specific focus in audio, speech, or multimodal machine learning.
Proficiency in Python and PyTorch for developing, maintaining, and debugging model evaluation and training code.
Demonstrated capability in writing reliable, maintainable code that scales beyond initial prototypes or notebooks.
Sound experimental judgment involving hypothesis formulation, baseline establishment, and result interpretation.
Demonstrated research and building capabilities through substantive open-source contributions, model releases, publications, or shipped systems.
Ability to iterate quickly and take independent ownership of open-ended technical challenges.
Clear verbal and written communication skills suitable for explaining technical limitations and methodologies.
Nice to Have
Background in training or fine-tuning transformer models.
Specialization in areas such as automatic speech recognition, text-to-speech, speech-to-speech, emotion recognition, prosody, or audio understanding.
Familiarity with preference learning, post-training, or generative model human evaluation.
Skills in optimizing the efficiency of large-scale data processing, distributed inference, or distributed training.
Prior experience working alongside external research groups on evaluation or model creation.
Skills and Tools
Python
PyTorch
Large-scale model training and fine-tuning frameworks
Distributed training and inference systems
Large-scale data processing workflows
Salary and Benefits
Annual salary range between $190,000 and $260,000 USD.
About the Company
Hume AI is a Series B startup headquartered at Hume AI HQ, focused on building AI optimized for human well-being and emotional understanding. The organization has released pioneering speech language models, conducts scientific research published in leading journals like Nature, and supports The Hume Initiative to promote ethical guidelines for empathic technology.
How to Apply
Interested candidates should visit the official Hume AI careers page to submit their application for the Research Engineer position.
