Job Summary
The Acoustics and Conservation Technology (ACT) Lab in the Applied Ocean Physics and Engineering Department at Woods Hole Oceanographic Institution is currently searching for a Research Assistant to work 3/4-time with a cross-disciplinary team focused on developing passive acoustic monitoring (PAM) applications to improve fisheries-independent stock assessments.
Areas of interest include automated sound detection and classification for ecologically and commercially important fish species (e.g., haddock, cod, herring), acoustic dataset curation and labeling workflows, and analysis of ambient noise to define and compute relevant noise metrics for PAM performance and interpretation. Significant duties include curating and quality-controlling large passive acoustic datasets, validating detector outputs, developing and evaluating machine learning methods for sound processing (including deep learning approaches such as convolutional neural networks and transformers), and testing detector performance and robustness across varying noise conditions. Familiarity with PAM workflows and tools and/or marine bioacoustics is required. Strong Python skills, ability to document code and workflows, and efficient execution in a collaborative research environment are required.
Job Description
This is an entry to early career level position designed to encourage the connection and application of academic training to results-oriented projects in support of scientific and research activities. Initially this position will have direct supervision to achieve structured and assigned objectives, and is expected to later expand to broader and more independent tasks. The candidate will be expected to work on tasks requiring creativity and independent thinking, along with a proven understanding of fundamental research principles and scientific computing practices. The candidate will have expertise in passive acoustic monitoring and machine learning for sound processing through either a recognized degree and relevant experience in a previous research or engineering position.
Essential Functions & Duties
- Curates, organizes, and quality-controls large passive acoustic monitoring (PAM) datasets, including associated metadata and documentation.
- Performs validation of detector outputs and contributes to creation of consistent annotation guidelines and workflows.
- Develops and applies signal processing and machine learning techniques for sound detection/classification, including deep learning approaches (e.g., convolutional neural networks and transformer-based models).
- Explores and implements unsupervised and/or self-supervised learning techniques for representation learning, clustering, anomaly/novelty detection, and/or reducing labeling requirements for detector development.
- Implements and maintains Python-based data processing and model training/evaluation pipelines (e.g., spectrogram generation, augmentation, training, inference, and benchmarking).
- Tests detector performance and robustness across deployments and varying acoustic conditions; conducts error analysis and recommends model and dataset improvements.
- Analyzes ambient noise conditions and defines/computes noise metrics relevant to PAM performance and interpretation (e.g., noise statistics that affect detectability).
- Contributes to software development best practices (version control, reproducible workflows) and produces clear documentation of methods, code, and results.
- Summarizes findings and presents results to the ACT Lab and project collaborators; supports preparation of tutorials and materials accompanying open-source software releases.
Required Experience & Education
- Electrical Engineering degree with minimal work experience. Experience working to develop and test ML/AI techniques in other fields will be viewed positively.
- Demonstrated proficiency in Python for scientific computing and data analysis.
- Demonstrated experience applying AI/ML techniques to audio/signal processing or related time-series data.
- Experience working with real-world datasets, including data QA/QC and clear documentation of methods and results.
Preferred Experience & Education
- Experience with passive acoustic monitoring (PAM) workflows and/or bioacoustics datasets; marine PAM experience is a plus.
- Experience with deep learning frameworks (e.g., PyTorch) and modern model architectures for detection/classification (CNNs, transformers).
- Experience with unsupervised/self-supervised learning for acoustic representation learning or low-label regimes (e.g., contrastive learning, masked prediction, clustering-based approaches).
- Familiarity with labeling tools/workflows, active learning or human-in-the-loop validation, and performance evaluation for detectors (precision/recall, PR curves, etc.).
- Experience with scalable computing (HPC/cloud), workflow automation, and/or containerized/reproducible environments.
- Experience contributing to open-source software (Git-based workflows, documentation, tutorials).
Additional Job Requirements
Hourly Rate: $[contact details available in the full listing] USD
The hourly rate provided for this position reflects the set base pay for new hires. Final level placement will be determined based on factors such as relevant skills, experience, and qualifications, as well as internal equity and market conditions.
WHOI accepts applications on a rolling basis - applications will be reviewed as they are received, and we encourage you to submit your application as soon as possible to ensure full consideration. While we will continue to review applications until the position is filled, and early applicants may have an advantage in the selection process.
Posted: 28-Feb-26
Location: Onsite - Woods Hole, MA
Type: Full-time
Categories: Academic/Faculty
Internal Number: JR100466
