Description
About the AI Division:
The AI Division is a unique group within Ceva, driving innovation in Machine Learning and Generative AI architectures for edge and cloud inference.
Our R&D spans Neural Network Processors (NPUs), Vision DSPs, and advanced AI solutions.
About the Role:
You will be a key contributor to Ceva’s AI Graph Compiler software stack for NPUs, designing system-level execution flows for advanced neural networks, including LLMs.
The role focuses on L1/L2 memory management, data movement, and performance optimization, working closely with compiler and hardware architects.
What will you do:
· Design and own key components of Ceva’s AI Graph Compiler.
· Develop and optimize neural network execution flows and memory management.
· Enable complex AI workloads, including LLMs, and implement new NPU features.
· Analyze performance bottlenecks and drive system-level optimizations.
· Collaborate with compiler and hardware teams on HW–SW solutions.
Requirements
Requirements:
· 5 years of software development experience using C/C++.
· BSc/MSc in Computer Science, Electrical Engineering, or equivalent.
· Experience designing and developing complex software systems.
· Strong understanding of memory management and performance optimization.
· Strong problem-solving skills and technical ownership.
Advantages:
· Experience with AI accelerators, NPUs, GPUs, or DSPs.
· Experience with AI compilers, graph optimization, or neural network execution.
· Experience with LLMs or other large neural network workloads.
· Familiarity with PyTorch, TensorFlow, ONNX, or Python.