True Anomaly
Software Engineer I, Perception (New Grad)
About this role
True Anomaly seeks a new graduate Software Engineer to develop hybrid perception systems for autonomous spacecraft, combining classical computer vision with deep learning to enable space-based object detection, tracking, and navigation. You'll implement tracking algorithms, train neural networks, optimize models for edge deployment, and validate end-to-end performance across simulation and hardware testing.
What you'll do
- Implement classical tracking algorithms including Extended Kalman Filters, Hungarian algorithm for data association, and track management logic
- Train neural networks for detection and classification using YOLO, ResNet, and learned appearance features for threat discrimination
- Build hybrid perception pipelines fusing neural network detections with classical tracking and coordinate transformations
- Develop image processing chains including hot pixel filtering, adaptive thresholding, centroiding, and star catalog matching
- Deploy and optimize models to edge hardware through quantization and integration with C++ inference engines
- Generate synthetic training data in Blender and validate performance through software/processor/hardware-in-the-loop testing
What they're looking for
- Python programming
- C++ (entry-level)
- PyTorch or TensorFlow
- Extended Kalman Filters and state estimation
- Computer vision and image processing
- Neural network training and optimization
- Coordinate transformations and linear algebra
- YOLO and object detection frameworks
Opens the application — the Jobs AI extension fills it for you. Set up autofill
Opens the official application on the employer’s site. No login required.
True Anomaly
True Anomaly develops advanced spacecraft and aerospace defense systems for mission-critical space applications. The company is hiring mechanical, thermal, systems, and test engineers to design, test, integrate, and verify complex space vehicle systems.
- Website
- trueanomaly.com
Likely interview questions
- Walk us through how you would implement an Extended Kalman Filter for multi-object tracking and what challenges you'd anticipate with occlusions or false detections.
- Describe your experience training neural networks in PyTorch—what datasets have you used and how did you handle class imbalance or domain shift?