Senior Android Computer Vision Engineer

Location: Budapest, Hungary
Ways of working: Hybrid
Contract: Permanent, full-time

About Hawk-Eye Innovations

Hawk-Eye Innovations is at the forefront of sports technology, delivering officiating, tracking, broadcast and performance solutions across some of the world's biggest sports. As part of Sony Sports, we continue to push the boundaries of innovation, helping athletes, coaches, teams, broadcasters and governing bodies make better decisions through technology.

Bring computer vision to the world of sport

What if athletes, coaches and teams could access the kind of tracking technology that has traditionally been out of reach?

We’re building a new generation of tracking services for elite and semi-professional sport, putting powerful computer vision technology directly into the hands of the people who need it most.

We’re looking for a Senior Android Computer Vision Engineer to help make that possible.

You’ll join a small, cross-functional engineering team building products at the cutting edge of sport and technology. With the direction set and plenty still to be built, you’ll have the autonomy to shape solutions, make important technical decisions and see your work make a real difference.

The role

You’ll own computer vision on the device, taking detection and tracking models and making them work reliably on consumer Android phones and tablets.

That means much more than getting a model to run. You’ll optimise and profile inference for real-world hardware, balancing performance with accuracy across different devices, lighting conditions and environments. You’ll work close to the camera capture pipeline, where frame timing, exposure and device behaviour can all affect the quality of the output.

Your challenge is to ensure that the data produced on a mobile device is accurate, reliable and trusted by the wider tracking system.

You’ll work closely with mobile, backend, computer vision and machine learning engineers, with the freedom to own both your technical approach and its implementation. We also embrace AI-augmented and AI-delegated development as part of how we work, enabling a small team to build and iterate at pace.

This is an opportunity to solve genuinely challenging engineering problems, influence how on-device vision is built and help establish the foundations for a technology that can change how sport is experienced and analysed.

What you’ll be doing

• Own on-device inference: Design and implement Android inference solutions that meet demanding accuracy, latency and performance requirements.
• Build the native Android vision layer: Develop and maintain C++ components using the Android NDK, alongside JNI bindings that connect them to our Kotlin application.
• Optimise models for mobile: Convert, quantise, profile and optimise detection and tracking models, exploring techniques such as pruning, graph optimisation and operator-level improvements.
• Solve real-world camera challenges: Work close to the capture pipeline, understanding how frame timing, exposure and camera behaviour influence model performance.
• Make accuracy measurable: Build validation tools and measurement processes that demonstrate performance against ground truth in real-world conditions.
• Engineer for real devices: Account for thermal constraints, battery consumption, hardware differences and the unpredictable conditions encountered in the field.
• Collaborate on model improvements: Work with our computer vision and machine learning specialists to identify, prototype and evaluate improvements to model performance.
• Shape technical decisions: Help determine what should run on-device and what should run elsewhere, bringing clear reasoning and an understanding of technical trade-offs.
• Build production-quality software: Write maintainable, testable and reliable code that meets the team's engineering standards.
• Work across disciplines: Communicate clearly with mobile application, backend and machine learning engineers who depend on your work.
• Build with AI: Use AI-assisted and delegated development tools as a standard part of your workflow, working within our security and intellectual property guidelines.