Assistant Research Engineer – AI/ML for Autonomous Systems

About the role

The University of Maryland MATRIX Lab seeks an Assistant Research Engineer to conduct applied AI/ML research for autonomous aerial, maritime, and underwater systems. This full‑time, on‑site position works closely with faculty and defense partners to translate machine‑learning advances into operational capabilities.

Key responsibilities

• Develop end‑to‑end AI/ML pipelines for perception, sensor fusion, and reinforcement‑learning‑based control on autonomous platforms.
• Design and implement adaptive, uncertainty‑aware learning methods such as knowledge distillation and Gaussian‑process‑based active context selection.
• Integrate multi‑modal sensors (RGB cameras, LiDAR, IMU, acoustic) and evaluate fusion architectures in the MATRIX Lab testbeds.
• Conduct controlled experiments in the OASIS and Hydrodynamics Lab, creating quantitative benchmarks for system performance.
• Deploy and maintain version‑controlled code, model packaging, and simulation‑to‑real transfer workflows on UAS, maritime, and underwater vehicles.
• Build interactive ML‑enabled tools and decision‑support applications using frameworks like FastAPI and React.
• Collaborate with interdisciplinary teams, prepare technical deliverables for government and industry stakeholders, and contribute to proposal writing.
• Mentor graduate and undergraduate researchers in AI/ML methods and experimental design.

Required profile

• Advanced coursework or research experience across multiple AI/ML subfields (computer vision, probabilistic modeling, control theory, etc.).
• Experience bridging software and hardware, including debugging onboard compute and sensor integration.
• Familiarity with safety‑critical or real‑time system constraints such as latency and fault tolerance.
• Proven ability to translate research outputs into tangible artifacts and communicate results to mixed audiences.

Required skills

• Python programming
• Machine‑learning frameworks: PyTorch, scikit‑learn
• Robotics middleware: ROS / ROS2
• Computer‑vision and multi‑modal perception techniques
• Reinforcement learning and policy optimization
• Sensor‑fusion architectures (RGB, LiDAR, IMU, acoustic)
• Knowledge‑distillation, Gaussian‑process active learning
• Web/API development: FastAPI, React
• Version control (git) and collaborative software practices
• Containerization and HPC/cloud environments for large‑scale model training