Engineering Intern - Crucible Synthetic NEI (Summer 2027)

Millennial Software Solutions • Chantilly, VA

4d • Onsite

About The Position

Millennial Software, part of Expedition Technology, is hiring two Engineering Interns for its Summer 2027 program at the company’s Chantilly, Virginia headquarters. Interns will work on Crucible, an agentic modeling and simulation platform, building a synthetic non-Earth imagery (NEI) generator from publicly available spacecraft models. The project addresses the scarcity and classification of real NEI by rendering spacecraft under varied pose, lighting, range, and sensor conditions, with per-component segmentation masks included for every frame. The generator is the primary deliverable and is expected to be complete by week six.

The stretch goal is to fine-tune a segmentation model on the synthetic dataset in Keystone and map predicted spacecraft components—such as the bus, solar arrays, antennas, thrusters, and apertures—into a populated Target Model Database (TMDB) record. All work is unclassified and uses publicly available models. Interns will be paired with mentors, participate in daily standups, and present a recorded demonstration and readout to leadership.

The program runs May 17 through August 6, 2027, for 12 weeks, full time and on site in Chantilly, VA, with remote flexibility. The final deliverable is a recorded pipeline run and leadership readout; targets are set by the mentor team. Compensation is $26.50–$27.50 per hour.

Requirements

• Be a U.S. citizen able to obtain and maintain a DoD security clearance.
• Be a rising senior, or an exceptional rising junior, pursuing a BS in Computer Science, Aerospace Engineering, or a related field, and meet the GPA requirement.
• Have strong Python skills and comfort with Git and Linux development environments.
• Have coursework or project experience in at least one of computer graphics or 3D rendering, computer vision, or machine learning (PyTorch or similar).
• Be available full time, on site in Chantilly, VA, for the 12-week program.

Nice To Haves

• Experience with rendering tools or engines such as Blender, Unity, Unreal, Three.js, or PyRender, or mesh formats such as glTF, OBJ, or STL.
• Image segmentation or object detection experience, such as U-Net, Mask R-CNN, or SAM.
• Orbital mechanics or spacecraft systems coursework.
• GPU training experience and familiarity with experiment tracking.

Responsibilities

• Build a rendering module in Crucible using open 3D spacecraft models and a parametric scene generator covering orbit geometry, target pose, illumination and self-shadowing, sensor range and resolution, noise, and blur.
• Emit per-component segmentation masks with every rendered frame and record rendering parameters so each image can be regenerated.
• Produce a labeled dataset of several thousand frames across multiple targets.
• Demonstrate dataset generation for a new target in under one working day.
• Stretch goal: Fine-tune a segmentation model based on a published architecture and run training and evaluation in Keystone against a held-out synthetic set.
• Stretch goal: Post-process predicted components into TMDB schema fields and populate one record end to end.
• Document the pipeline and present at the leadership readout.

What This Job Offers

• Job Type: Full-time
• Career Level: Intern