We are looking for an AR Algorithm Engineer to develop perception and tracking algorithms for real‑world defense applications, with a focus on estimation problems (pose, motion, structure) in GPS-denied environments.
Our workflow is science-driven: rapid Python prototyping to explore ideas and failure modes, followed by rigorous validation (experiments, regression tests), and finally C++ implementation for embedded deployment. Early work is intentionally fast-moving and exploratory; rigor increases as solutions mature. You’ll work in a collaborative team of experienced, academically grounded engineers where discussion, reproducibility, and shared understanding matter.
Responsibilities
- Develop algorithms for state estimation, SLAM/VIO, sensor fusion, and tracking
- Prototype in Python, design experiments, and build regression tests
- Work with noisy real-world sensor data (cameras, IMUs, etc.)
- Optimize and deploy in C++ on embedded platforms (Jetson, Qualcomm, Android)
- Collaborate closely with teammates to review, validate, and improve approaches
- Strong foundations in estimation & optimization (Kalman filters, smoothing, nonlinear least squares / factor graphs)
- Experience with SLAM, VIO, or 3D vision systems
- Solid linear algebra, probability, and numerical methods
- Comfortable with a Python-first prototyping workflow
- Ability to implement in modern C++ for performance-critical systems
- Experience with real sensor data and understanding noise/bias/failure modes
- Comfortable in a Linux + git development environment
- Strong collaboration skills; able to work effectively in a research-oriented team
Stack• Python, C++, ROS2, Eigen• Linux, git
Nice to have: Ceres, GTSAM, embedded/mobile deployment, calibration & sensor modelingNot a Fit If
- You prefer going straight to C++ without prototyping and validation
- You avoid collaborative technical discussion or experiment-driven work
