Frontier Audio, based in Chicago, is a stealth-mode startup developing a game-changing wearable/hearable device that will make tens of millions of critical frontline workers. The company has raised funding from the first investor in Oculus (AR/VR company sold to Meta), contracts with the United States Department of Defense, pilots scheduled with Fortune 500 companies, and an ambitious team. The company is seeking those who want to overcome huge technical challenges and do the most impactful and important work of their life delivering long overdue capabilities to a huge, underserved user group.
Job Description |
Frontier Audio Labs is looking for an extremely passionate and hardworking speech AI engineering intern with experience developing audio algorithms. The selected individual will contribute to key parts of the R&D and product development process and their work will have a huge impact on the safety, comfort, and productivity of millions of end-users powering and protecting the world. "Crazy ideas", being scrappy, and having fun are tenants of the culture as well as focus and hard work, fundamental requirements to reach our goals. Frontier Audio is an ideal company for those looking to be a cornerstone or foundational member of a quickly growing team that iterates rapidly. While Frontier Audio as a startup may not be able to offer top-bucket cash compensation starting off, this position will include a significant (multiple percent) equity stake.
Responsibilities |
Conduct research and development on cutting-edge speech and audio AI models for a full duplex speech interface on a hearable/wearable device
Design and implement large-scale audio foundation models and specific speech and audio models for speech recognition, text-to-speech synthesis, acoustic event understanding, user detection, etc.
Create natural language understanding (NLU) systems for intent recognition and context awareness and dialog management systems for smooth turn-taking, interruption handling, and error recovery in user-device conversations
Leverage cloud-based LLMs or device-based LLMs trained at second level on user specific database
Rapidly prototype and subjectively and empirically test
Optimize speech models and processing pipelines for real-world applications, focusing on accuracy, latency, power efficiency, and edge compute constraints
Work with large-scale speech and audio datasets to improve model performance and scalability
Collaborate with audio algorithm and audio system R&D team as well as hardware and embedded software engineers
Qualifications |
Required
Dedication to creating products stretching the boundaries of possible
Understanding of state-of-the-art speech and audio foundation models, including ASR, TTS, dialogue systems, and audio modality LLMs
Strong programming skills in Python, C++, or C and experience with deep learning frameworks (PyTorch, TensorFlow)
Strong foundation in machine learning principles and algorithms, particularly for real-time speech and audio applications
Hands-on experience implementing and optimizing deep neural networks for speech/audio tasks, ideally at the edge
Experience with large-scale speech/audio datasets and distributed training pipelines
Must have ability to be qualified to work in the U.S. in less than 3 months
Ability to commute to Frontier Audio's Chicago office on a daily basis
Preferred
Experience implementing full-duplex speech interfaces or spoken dialogue systems leveraging cloud/edge LLMs
Research experience with foundation models and generative AI specifically for speech and audio applications
3+ years of experience with speech processing tasks (ASR, TTS, speaker diarization, wake word detection) and natural language understanding
Familiarity with audio dataset curation and model evaluation
Familiarity with real-time audio processing pipelines and system architecture for wearables
Experience optimizing models for edge deployment (quantization, pruning, distillation)
Proven ability to drive research projects from conception to completion
Members of our engineering team have advanced degrees, but if you have taken a different route to gain the required experience, that is acceptable
The applicant must be a flexible, results-oriented problem-solver who is passionate about introducing large-scale innovation into the audio industry.
Frontier Audio Labs evaluates qualified applicants without regard to race, color, religion, sex, sexual orientation, gender identity, genetic information, national origin, age, disability, veteran status, or any other legally protected characteristics.
All inquiries should be directed to pmoeckel@frontieraudio.com.