AI Compiler Engineer

The opportunity

We are working with an international semiconductor company developing high-performance AI processors for intelligent edge systems.

They are looking for an AI Compiler Engineer to build the software layer that enables machine-learning models to run efficiently on custom accelerator hardware. You will work across model frameworks, compiler infrastructure, runtime software and processor architecture, solving the problems that sit between an AI model and its execution on silicon.

The role is suitable for someone with strong C++ and systems fundamentals who enjoys understanding how software maps onto hardware. Applications are open to both experienced compiler engineers and high-quality graduates with relevant research or project work.

• Build and improve compiler technology for custom machine-learning accelerators
• Translate neural-network models into efficient hardware-executable workloads
• Develop compiler passes covering graph transformation, operator lowering, scheduling and code generation
• Optimise computational graphs through techniques such as operator fusion, constant folding and layout transformation
• Improve the use of on-chip memory, data movement and parallel compute resources
• Develop scheduling strategies that account for bandwidth, latency and hardware dependencies
• Support model ingestion from frameworks and formats such as PyTorch, Tensor Flow, ONNX and related ecosystems
• Create mappings for neural-network operators onto specialised processing units
• Analyse workloads to identify compiler, runtime and hardware bottlenecks
• Debug functional accuracy, numerical behaviour and performance issues across the compilation pipeline
• Benchmark models using simulation environments, development platforms and production silicon
• Work closely with processor architects to enable new instructions and hardware capabilities
• Collaborate with kernel, runtime, driver and AI framework engineers
• Build testing, validation and automation tools for compiler development
• Support the deployment of production AI workloads onto embedded and edge-computing platforms

• A Master’s degree in Computer Science, Computer Engineering, Electrical Engineering or a related technical discipline
• Strong software development skills in C++
• A solid understanding of algorithms, data structures and software architecture
• Knowledge of compiler design, intermediate representations or code-generation concepts
• An understanding of computer architecture and how software is executed by hardware
• Experience debugging complex software across multiple system layers
• Ability to work effectively within a large production codebase
• Strong analytical skills and an interest in performance engineering