Audio AI Research Scientist — Audio Understanding

Bose Corporation

2d•$165,400 - $227,500

About The Position

Bose Corporation is seeking an Audio AI Research Scientist to develop novel AI-powered audio processing algorithms for real-time operation on physical devices. The work supports applications such as context understanding, dynamic awareness, and augmented hearing. The scientist will collaborate with experts in machine learning, digital signal processing, software engineering, and psychoacoustics to prototype and implement algorithms. The role focuses on audio understanding algorithms that enable devices to recognize events and conditions in their surroundings. Areas of interest include sound event detection and localization (SELD), acoustic scene understanding, unusual or anomalous sound detection, and multimodal context understanding using audio, video, and motion sensors. The position spans research and experimentation through evaluation, integration, and real-time device deployment.

Requirements

Graduate degree (MS or PhD) in machine learning, computer science, electrical engineering, or a related field, or equivalent research experience.
Strong hands-on experience with modern audio understanding models, including sound event detection, SELD, acoustic scene analysis, or anomaly detection.
Ability to tackle complex, open-ended problems through experimentation and iteration.
Experience developing and evaluating deep learning models in PyTorch.
Strong Python skills.
Understanding of audio representations, DSP, and evaluation metrics for detection, localization, and generalization.
Strong communication skills.

Nice To Haves

Experience with multimodal fusion using video, IMU, or other sensors for context understanding.
Experience with efficient inference, causal ML, ONNX/TFLite, embedded ML, or C/C++.
Conference publications such as ICASSP, Interspeech, NeurIPS, or similar are welcome but not required.

Responsibilities

Research, prototype, train, and evaluate novel audio understanding models from early experiments through real-time deployment on devices.
Develop robust methods for sound event detection and localization, acoustic anomaly detection, and context-aware audio interpretation.
Explore multimodal approaches combining microphone signals with video, IMU, or other sensor data when useful.
Build and curate datasets and data pipelines, evaluation protocols, and demonstrations measuring accuracy, robustness, latency, and compute cost.
Collaborate with DSP and software engineers to integrate models into proof-of-concept systems and product platforms.
Communicate results and contribute to patents or conference papers.