Compiler Engineer (Neural Engine)

November 5, 2026

Office

Lead

München, Germany

Indeed

Keep in mind

Office-only work

Also worth knowing

English is required

Description

The software team develops the compiler and runtime for machine learning accelerators on custom silicon. The role involves analyzing, designing and implementing hardware-interfacing software for AI applications across a wide range of product platforms, plus compiler optimizations to improve latency and bandwidth of ML models. Development spans compiler, framework, driver and firmware layers, with performance profiling, collaboration with model teams and engagement with QA

Responsibilities

Analyze, design and implement software interfacing with hardware for AI applications
Assume technical leadership, mentor and guide other developers
Ensure technical excellence and robust, maintainable software
Design and develop compiler optimizations to improve latency and bandwidth of ML models
Develop software across compiler, framework, driver and firmware layers
Collaborate closely with model teams to support new ML models on hardware
Conduct performance profiling and diagnostics using and developing specialized tools
Engage with QA teams during development

Requirements

Bachelors degree in an aligned field and some industry experience
Hands-on experience with compiler development
Excellent programming skills in C++, C or Swift
Proficiency in English for professional work
Strong written and verbal communication skills
MS or PhD in computer science, machine learning or related field
Experience with low level software architecture, modeling or validation
Familiarity with AI networks such as CNN, transformer and diffusion architectures
Proficiency in profiling and optimizing complex software
Experience with SoC low level software development, distributed computing or hardware acceleration
Strong background in data science and statistical methods analyzing large datasets
Experience with the MLIR compiler framework
Experience building debugging and performance analysis tooling