Staff AI Platform Engineer — Heterogeneous Compute

Axelera AI; Polska; Remote; PLN 528,000 - 705,000; Full time

Benefits offered by this job

pension plan

extensive employee insurances

the option to get company shares

Job summary

Axelera AI is building the next-generation AI platform to support anyone who wants to help advancing humanity and improve the world around us. We are a fast-growing deep-tech company with a global presence and offices across Europe, including Leuven, Amsterdam, Eindhoven, Zurich, Florence, Milan, and Bristol, plus remote opportunities across Europe.

Staff AI Application Engineer in the Model Enablement team will architect and implement the model enablement platform, decide model runtimes across

Qualifications

• 8+ years of engineering experience in systems, compilers, runtimes or performance.
• Hands-on experience with heterogeneous computing across CPU, GPU, DSP, NPU or FPGA.

Responsibilities

• Build the model enablement platform and agent harness to partition, compile, run and profile models.
• Capture, analyze and split models across AIPU units, host CPU and accelerators with placement tags.
• Extend split to large models and multi-model pipelines with caching and offloading.
• Collaborate with graph compiler and runtime teams to optimize end-to-end performance.
• Use AI coding agents to assist in building, testing and verifying work.

Skills

C++ programming

Python

Computer architecture

Parallel programming

Heterogeneous compute

AI coding agents

Tools

MLIR

ONNX

Torch.export

TVM

CUDA/OpenCL/Vulkan

Job description

Axelera AI is building the next-generation AI platform to support anyone who wants to help advancing humanity and improve the world around us. We are a fast-growing deep-tech company with a global presence and offices across Europe, including Leuven, Amsterdam, Eindhoven, Zurich, Florence, Milan, and Bristol, plus remote opportunities across Europe.

Staff AI Application Engineer in the Model Enablement team will architect and implement the model enablement platform, decide model runtimes across