Data Scientist II - Machine Learning for Undersea Acoustics

In PNNL's AI and Data Analytics Division (National Security Directorate), the data scientist investigates and develops deep learning methods for undersea acoustics: analysing acoustic data and building, training and evaluating models that detect, classify, localise, separate or characterise underwater sources and phenomena for national-security and naval-sensing missions. Work combines deep learning with Fourier analysis and signal processing, designs reproducible experiments comparing detection performance, false alarms, robustness to propagation and environmental variability, uncertainty and computational cost, and extends to time-series analysis, representation learning, anomaly detection, uncertainty quantification, multimodal sensing and evaluation of agentic AI systems that combine deep learning with acoustic analysis tools.

The salary range is $114,000-$182,100 per year.

BS/BA with two years of data-science experience, or MS/MA, or PhD; experience developing, training and evaluating deep learning models for acoustic, sonar, audio, radar, seismic or related sensor data; implementing and training models in Python with PyTorch. Preferred: undersea acoustic detection/classification work, Fourier or signal-processing methods, time-frequency or array-processing methods, and agentic AI evaluation. The role requires an onsite presence in Richland or Seattle.