NIWC - Undergraduate Student III - AIRED Project (Task No. 56-182608)

SDSU Research Foundation • San Diego, CA

1d • $20 • Onsite

About The Position

The hourly rate for this position is $19.67 per hour for Undergraduate III and is non-negotiable. This position is open until filled Under direction, incumbents must have course work in the discipline requested on the task order Statement of Work. This Statement of Work covers the need for student services to support research and development activities related to reinforcement learning (RL) algorithms for modeling adaptive target behaviors. Under the guidance of the engineering team, the student will contribute to the design, implementation, and evaluation of RL agents and simulation environments. Work will include developing custom RL environments using Gymnasium, implementing and training agents with Stable Baselines3, and assisting in the analysis of agent performance within active sonar operational scenarios. The student will participate in iterative experimentation, documentation of methods and results, and preparation of technical materials for potential publication or presentation. Tasks will be scoped to match the student s academic background and skill level, with opportunities to gain experience in applied machine learning, simulation development, and defense-focused AI research. The AIRED project focuses on researching and developing machine learning-based technical solutions to advance the Navy s ability to model intelligent, adaptive target behaviors in active-sonar Anti-Submarine Warfare (ASW) environments. Today s decision-support tools primarily emphasize acoustic performance prediction and often lack representations of reactive, tactically aware target decision-making. AIRED aims to address this gap by creating reinforcement learning (RL) agents that can learn and demonstrate realistic, adaptive behaviors in simulated operational scenarios. By building models that behave more like real targets, the project will help improve the realism of training tools, support the development of new tactics and sensor strategies, and ultimately strengthen the Navy s ability to plan and operate in complex undersea environments.

Requirements

• Undergraduate Student 3: Must have course work in the discipline requested on the task order Statement of Work.
• The student must be at the Senior level and have experience in 100% of the task requirements and be able to accomplish 90% of the work independently.
• Is not required to be an SDSU Student.
• Undergraduate 3: Computer Science major preferred; relevant majors such as Data Science or Cognitive Science will be suitable given relevant coursework and/or work experience.
• Required Coursework and/or relevant work and/or laboratory experience: Machine Learning Algorithms Programming Design & Analysis of Algorithms Data Structures Data Science

Nice To Haves

• Desired Coursework and/or relevant work and/or laboratory experience: Neural Networks/Deep Learning Linear Algebra Probability
• Python software programming experience.
• Experience with GIT and version control.
• Experience in reinforcement learning.
• Ability to read and implement technical documents or test standards.
• Comfortable with working independently and in group settings.
• Independent/self-starter who can operate under minimal oversight.
• Collaborative/team-player who can work well with groups.

Responsibilities

• Assist with the design, development, and sustainment of custom RL environments using Gymnasium.
• Assist with implementation, training, and optimization of RL agents using the Stable Baselines3 framework.
• Develop Python-based software modules with strong adherence to coding standards, documentation practices, and reproducible experimentation workflows.
• Apply core RL concepts, including policy optimization methods (e.g., PPO), action and state space design, and reward shaping, to support model development and evaluation.
• Utilize Stable Baselines3 and Gymnasium to integrate, test, and refine RL algorithms within simulation environments.
• Use Git for version control, including branching, merging, and maintaining clean commit histories to support collaborative development.
• Demonstrate effective problem solving skills by identifying issues, proposing solutions, and iteratively improving model performance and software reliability.
• Engage proactively in learning new tools, frameworks, and methods relevant to machine learning, simulation development, and reinforcement learning research.

What This Job Offers

• Job Type: Part-time
• Career Level: Intern
• Education Level: No Education Listed
• Number of Employees: 501-1,000 employees