Computer Science & Computer Vision AI/ML Intern

ITERRA is building a passive, edge-native sensing system that detects and classifies low-altitude unmanned aircraft systems in denied and degraded environments. The Computer Science & Computer Vision AI/ML Intern will develop and evaluate computer vision models, deploy them on NVIDIA Jetson hardware, build datasets and tooling, collect field data, and integrate electro-optical outputs with radio-frequency sensing.

Responsibilities:

• Computer vision on the edge. Detection and tracking of Group 1-3 airframes in visible and infrared imagery. Training, evaluation, and the honest confusion matrix that comes with them. Small, fast, low-power models, not leaderboard models. The electro-optical and infrared line is a development track, not a shipped product, and you would be working on it from the beginning
• Deployment on NVIDIA Jetson. Getting a model off your laptop and onto the unit: quantization, TensorRT, throughput and latency measured on the actual board under Caribbean heat, not estimated
• Dataset and labeling pipeline. Collection, labeling protocol, versioning, augmentation, train and validation splits that do not leak. The training corpus is a company asset and you help build it
• Field data capture. You come to the range. Live airframes, fixed and moving, multiple types. You help run the collection and you carry the gear
• Evaluation you would defend. Precision, recall, false alarm rate, and the conditions each was measured under. Written so it can drop into a federal proposal without rewriting
• Fusion with the RF side. Working with the RF engineer so that what the antenna detects and what the camera sees agree on a single track. Both halves are passive. The system listens and looks; it never transmits
• Tooling. The small scripts, dashboards, and harnesses that make the above repeatable by someone who is not you