Computer Vision Engineer

About the role

You will architect and implement core perception modules that enable real-time decision-making for autonomous platforms in contested environments. This role owns the design of computer vision algorithms from initial concept through to high-performance deployment on embedded hardware. You will be responsible for translating abstract product requirements into robust, low-latency C++ vision pipelines that meet strict operational standards. You will collaborate closely with AI researchers and systems engineers to ensure that perception components align with end-to-end autonomy goals. The position requires a proactive approach to debugging complex interactions between sensor data, software architecture, and hardware constraints. You will contribute to the creation of synthetic training data strategies that improve model robustness without relying solely on real-world collection. This role demands ownership of the full software lifecycle, from initial prototype to safety-critical validation in simulation and flight. You will play a key role in defining best practices for code quality, testing, and documentation across the perception team.

Key facts

What you'll do

Architect and maintain high-throughput C++ pipelines for image processing and feature extraction that operate under strict latency budgets.

Design and implement modules for sensor calibration, image rectification, and multi-camera synchronization to support accurate 3D reconstruction.

Develop real-time detection, classification, and tracking algorithms that function reliably in diverse lighting and weather conditions.

Integrate classical computer vision techniques with modern machine learning models to create hybrid perception systems.

Optimize memory usage and compute efficiency to ensure algorithms run effectively on resource-constrained edge devices.

Conduct rigorous performance testing and validation to verify that vision modules meet safety and reliability requirements.

Work with data engineering teams to curate and annotate diverse datasets that improve model generalization across operational scenarios.

Collaborate with simulation engineers to create virtual test environments that accelerate algorithm development and regression testing.

Implement interfaces to communicate perception outputs with higher-level planning and control software components.

Lead technical investigations when facing difficult problems in areas such as low-light navigation, dynamic object detection, or sensor degradation.

Document technical designs and implementation details to ensure knowledge transfer and maintainability of the codebase.

Support on-site testing campaigns and contribute to the analysis of field data to drive iterative improvements in perception accuracy.

Mentor junior engineers by providing code reviews, technical guidance, and clear explanations of complex vision concepts.

Champion continuous improvement by refactoring legacy code, updating dependencies, and adopting modern C++ standards.

Participate in the definition of product-level requirements to ensure that technical solutions are feasible and aligned with customer needs.

Requirements

Demonstrated professional experience writing high-performance C++ code for real-time or embedded systems applications.

Strong proficiency in computer vision fundamentals, including but not limited to feature detection, matching, geometric vision, and camera models.

Experience with popular open-source computer vision libraries such as OpenCV, including both traditional and machine learning-based approaches.

Solid understanding of deep learning concepts and familiarity with deploying neural networks in production environments.

Proven ability to work effectively in a fast-paced, deadline-driven defense or aerospace environment.

Excellent problem-solving skills and a methodical approach to debugging complex software systems under time pressure.

Clear written and verbal communication skills to collaborate with multidisciplinary teams and explain technical trade-offs to non-technical stakeholders.

Eligibility to obtain security clearance as required by Australian government regulations for defense-related work.

Nice to have

Experience with sensor fusion pipelines that combine vision with other modalities such as radar or inertial measurement units.

Knowledge of machine learning frameworks such as TensorFlow or PyTorch, and experience with model optimization techniques like quantization or pruning.

Familiarity with simulation tools and game engines used for synthetic data generation and virtual testing.

Background in unmanned aerial systems or prior exposure to drone, VTOL, or similar platforms.

Experience with version control workflows using Git and CI/CD practices for automated testing and deployment.

Practical notes

This role is based in Port Melbourne, Australia, and requires availability for standard working hours during business days.

Travel may be required between office locations and to customer sites as necessary.

Employment is subject to relevant security vetting and clearance requirements.

Candidates must meet Shield AI's eligibility criteria as outlined in the requirements section to be considered for this position.