There is a category of work that humans cannot safely do. Inspecting the inside of a nuclear reactor. Surveying a disaster zone in real time. Assembling composite parts at tolerances measured in microns, for hours, without fatigue. Performing surgery through a 1-cm incision.

The answer is robots — and the field building them is one of the fastest-growing career areas in STEM. The 2025 robotics market crossed $100 billion globally. The specialized workforce to fill it doesn't yet exist.

What the Labs Are Doing

Idaho National Laboratory — Nuclear Hot Cell Manipulation

INL is the DOE's lead laboratory for nuclear energy research — and that means robotic systems are central to almost everything they do. INL operates the world's most advanced nuclear robotic testbeds, including remote manipulation systems for spent nuclear fuel handling and automated inspection rigs for reactor pressure vessels.

Their Advanced Robotics in Radiological Environments program pairs computer vision with force-feedback manipulators that can sense and respond to material properties without direct human contact. INL is also a leader in nuclear human-robot teaming — designing the interaction protocols that let a human operator in a control room guide a robot arm through a task it has never seen before, in a room too radioactive to enter.

Oak Ridge National Laboratory — Manufacturing Demonstration Facility

ORNL's Manufacturing Demonstration Facility (MDF) is DOE's flagship advanced manufacturing testbed. The MDF's robotic systems aren't just production tools — they are research instruments. ORNL's Collaborative Robotics and Additive Manufacturing group studies how autonomous robotic cells can adapt their fabrication strategy in real time based on sensor feedback, closing the loop between design, deposition, and quality verification.

Their work on robotic additive manufacturing of nuclear components — where material properties must be characterized mid-build, not just afterward — is pushing the limits of what in-process autonomy can achieve. ORNL also operates robotic systems for neutron scattering experiments, where autonomous sample handling dramatically increases throughput at instruments that are always oversubscribed.

Pacific Northwest National Laboratory — Hanford Site AUVs

PNNL is at the center of the most consequential environmental cleanup operation in American history: the Hanford Site in eastern Washington, where 177 underground tanks hold 56 million gallons of radioactive waste. Characterizing what's in those tanks — without opening them, without risking worker exposure — requires robotic systems that can navigate confined spaces, perform chemical sampling, and transmit high-fidelity data through concrete walls.

PNNL's Remote Systems and Robotics group has developed miniaturized crawler robots, tethered inspection probes, and autonomous underwater vehicles (AUVs) for in-tank waste characterization. This work directly informs decisions about a $100+ billion cleanup program.

Sandia National Laboratories — Assured Autonomy

Sandia has one of the longest robotic R&D histories in the DOE complex, rooted in national security applications. Their current frontier is assured autonomy — building robotic systems that can operate reliably in contested or degraded environments where GPS is unavailable, communications are jammed, and the environment doesn't conform to prior maps.

Sandia also does substantial work in robotic nonproliferation — automated systems for materials accountancy and safeguards monitoring at nuclear facilities worldwide. This is one of the few robotics specializations that requires deep knowledge of both robotics and nuclear policy.

Argonne National Laboratory — Polybot and Chain-of-Thought Autonomy

Argonne developed Polybot, an autonomous chemistry laboratory: it designs an experiment, executes it using robotic liquid handling, reads the results with integrated spectrometers, and uses active learning to decide what to test next — all without human involvement between cycles. Argonne's Chain-of-Thought Autonomy program explores how AI planning layers can be coupled with physical robotic systems for multi-step scientific tasks.

Career Tracks Covered in This Issue
  • Robotics/Autonomy Engineer — ROS 2, MoveIt, Gazebo, Isaac Sim. At national labs, add radiation-hardened hardware constraints and fail-safe requirements. Entry-level typically requires B.S. in ME, EE, or CS with a robotics project portfolio.
  • Controls Systems Engineer — Designing and tuning controllers for manipulators, mobile platforms, and actuated systems under model uncertainty. Strong mathematics required (control theory, dynamics, signal processing). Community college students with mechatronics coursework have a realistic CCI → SULI → permanent path.
  • Computer Vision / Perception Researcher — 3D reconstruction, anomaly detection in inspection imagery, force estimation. Heavy emphasis on robust perception under degraded conditions — low light, particulate contamination, radiation-induced sensor noise. PyTorch + OpenCV baseline.
  • Human-Robot Interaction (HRI) Researcher — High demand at DOE labs, where robots routinely operate as tools for human experts. Designing the interfaces and shared-autonomy protocols that let a domain expert guide a robotic system through complex tasks they understand but can't physically perform.

The Core Skill Stack

ROS 2
Python / C++
OpenCV
PyTorch (CV)
Control Theory
Gazebo / Isaac Sim
Linux / Bash
Mechatronics

How to Get There

SULI offers placements specifically in robotics and autonomous systems at INL, ANL, PNNL, ORNL, and Sandia. Search the SULI portal for "robotics," "autonomous systems," "remote systems," or "manufacturing" to find active projects. Applications open twice a year.

CCI is particularly well-suited for students with mechatronics, electronics, or computer networking backgrounds from community college programs. Robotics technician roles at labs — maintaining, calibrating, and operating robotic systems — are a legitimate entry point that can transition to research roles over time.

What to build before you apply:

  • A ROS 2 project, even a simple one. A simulated mobile robot navigating a maze is proof you understand the framework.
  • Any computer vision project using OpenCV or a PyTorch-based detector. A custom dataset and trained model is better than a tutorial.
  • Evidence that you can work in hardware and software simultaneously. A small mechatronics build — even an Arduino-controlled servo rig — demonstrates the integration mindset that lab robotics requires.

The national labs are building systems that will handle radioactive materials, characterize century-old waste, and accelerate scientific discovery at a pace human researchers alone cannot match. The engineers who build those systems are working now. The interns who will become those engineers are applying to SULI this year.

Resources to Go Deeper

  • DOE SULI Applications — For undergraduates. Target: INL Nuclear, ORNL MDF, PNNL Applied Physics, Sandia Robotics.
  • DOE CCI Applications — For community college students with mechatronics, electronics, or CS backgrounds.
  • ROS 2 Documentation — Start here. Free, comprehensive, the standard framework for every serious robotics research group.
  • INL Robotics Programs — Search "remote systems" or "robotics" on inl.gov for current research areas and student opportunities.
  • PNNL Remote Systems — Search "robotics" on pnnl.gov for the Remote Systems and Robotics group's current work.