Embedded Software Engineer GPU & edge AI Signal Processing

As Embedded Software Engineer for accelerated computing you bring our real-time sensor processing chain to life on state-of-the-art embedded GPU compute modules: you implement and optimize high-throughput signal processing algorithms for massively parallel execution, build low-latency streaming pipelines from the digital front-end into the processing chain, and continuously push the limits of what modern edge hardware can deliver - with data rates of several GB/s and hard latency budgets.

• Algorithm implementation & optimization

• Implement signal processing algorithms (FFT-based processing, filtering, detection and estimation stages) as highly parallel GPU kernels
• Optimize for throughput and latency: memory hierarchy, kernel fusion, asynchronous streams/graphs, mixed precision

• Streaming & data movement

• Design zero-copy / DMA-based ingest paths from FPGA-based front-ends into GPU memory
• Build deterministic, real-time-capable streaming pipelines including buffering, backpressure handling and precise time-stamping

• Platform & framework work

• Evaluate and adopt state-of-the-art vendor SDKs, libraries and pipeline frameworks for accelerated sensor processing and edge AI inference
• Maintain OS images, board support and deployment for the embedded compute modules

• Quality & verification

• Profile and benchmark systematically; validate numerically and bit-true against Python/C++ reference implementations
• Contribute unit/integration tests, CI/CD integration and documentation

• Completed technical studies, e.g. in computer science or electronical engineering
• Strong modern C++ (C++17 or newer) on Linux
• Solid hands-on CUDA / GPU computing experience: kernel development, profiling and performance optimization
• Understanding of real-time data streaming and heterogeneous memory architectures (DMA, pinned/unified memory, peer-to-peer transfers)
• Fundamentals of digital signal processing
• Experience with embedded GPU / edge-AI compute modules and their SDK ecosystems advantageous
• Experience with graph- or pipeline-based streaming frameworks for sensor or media processing advantageous
• Python for prototyping, validation and tooling advantageous
• Experience with high-rate sensor data (radar, lidar, imaging, software-defined radio) advantageous
• Strong team and communication skills in a small, interdisciplinary prototyping team
• Analytical and structured way of thinking and working
• Self-directed, proactive working style with a high level of ownership
• Precision and quality awareness, especially in verification and interface definition
• Pragmatic, hands-on prototyping mindset - willingness to iterate fast and learn from hardware
• Willingness to travel for integration and measurement campaigns (~15%)
• Very good German and English skills

Thales in Germany is part of an internationally leading technology group in the defence and security, aerospace as well as cyber security and digital identity markets., The Group invests more than €4,5 billion per year in Research & Development in key areas, particularly for critical environments, such as Artificial Intelligence, cybersecurity, quantum and cloud technologies. In 2025, the Group generated sales of €22.1 billion. For our more than 85,000 employees in 65 countries we open up visionary perspectives, realise individual career paths and enable creative freedom. This is achieved with courage, versatility and the firm intention to make the demanding challenges of our time safer and more inclusive. With our sustainable value-focused management we support diversity actively. Say HI* - Your journey to us At times of change our international teams are ready to meet the complexity of today with the industry-leading technologies of tomorrow. Will you be part of it? Your Talent Acquisition contact Maria Gaissert is looking forward to your online application.