AI/Machine Learning Research Scientist – Undersea Acoustics & Physics-Based Modeling

Peraton•Stennis Space Center, MS

4d•$104,000 - $166,000•Onsite

About The Position

Peraton is seeking an AI/Machine Learning Research Scientist to conduct basic and applied research in machine learning, deep learning, generative AI, large language models, and physics-based machine learning. This onsite position at Stennis Space Center will develop and evaluate advanced computational methods for Navy-relevant scientific, engineering, acoustic, and decision-support challenges. The work focuses primarily on underwater acoustics and acoustic propagation, with potential applications in wave physics, signal processing, and related computational-physics domains. The scientist will develop rapid machine-learning surrogate models for computationally demanding physics models, preserving decision-relevant fidelity while improving computational efficiency and supporting reliable technical analysis. The role also involves multidisciplinary collaboration, controlled experiments, model evaluation, reproducible research software, and communicating results through technical reports, publications, briefings, and presentations.

Requirements

Minimum experience: 8 years with a bachelor's degree, 6 years with a master's degree, or 3 years with a Ph.D.
Ability to obtain and maintain a Secret security clearance.
U.S. citizenship.
Strong analytical, problem-solving, technical communication, and quantitative research skills.
Demonstrated experience applying machine-learning or deep-learning methods to scientific, engineering, physics-based, or computational modeling problems.
Experience developing, training, testing, and evaluating machine-learning models using a modern framework such as PyTorch, TensorFlow, or JAX.
Knowledge of core machine-learning concepts, including neural-network optimization, regularization, representation learning, generative modeling, data management, model evaluation, and reproducible experimentation.
Ability to formulate research problems, design controlled experiments, analyze results, identify model limitations, and communicate defensible technical conclusions.
Ability to work independently while collaborating effectively with multidisciplinary technical teams and operational stakeholders.
Ability to develop and maintain reproducible research software and documentation.

Nice To Haves

Active Secret security clearance or higher.
Experience with underwater acoustics, acoustic propagation, wave physics, signal processing, or related computational-physics applications.
Experience with physics-informed machine learning, including physics-informed neural networks, neural operators, surrogate modeling, reduced-order modeling, or hybrid physics-machine-learning approaches.
Experience developing surrogate models or emulators that accelerate computationally intensive physics-based simulations.
Familiarity with uncertainty estimation, reliability assessment, and evaluating model generalization under distribution shift.
Experience applying generative AI or large language models to scientific, engineering, or decision-support challenges.
Experience collaborating with acousticians, physicists, engineers, scientists, or other multidisciplinary research teams.
Ability to rapidly learn new scientific domains and translate domain-specific requirements into effective AI/ML workflows.

Responsibilities

Collaborate with acousticians, physicists, engineers, materials scientists, computer scientists, and other subject-matter experts to develop and improve machine-learning approaches for complex technical problems.
Conduct research in deep learning, generative AI, large language models, and scientific machine learning for acoustic and physics-based applications.
Develop, evaluate, and implement physics-informed neural networks, neural operators, surrogate models, hybrid physics-machine-learning approaches, and generative models.
Design and execute controlled experiments to train, test, and evaluate machine-learning models.
Assess model accuracy, robustness, reliability, physics consistency, and generalization to conditions not represented in training data.
Develop machine-learning surrogates and other accelerated computational approaches that reduce the cost of full-physics simulations while maintaining fidelity to decision-relevant outputs.
Translate physics and engineering requirements into practical machine-learning approaches, data representations, performance metrics, and validation protocols.
Develop and maintain clear, reproducible research software, technical documentation, and experimental records.
Analyze quantitative results, identify model limitations, uncertainty, and reliability considerations, and communicate technically sound conclusions through reports, publications, briefings, and presentations.