AI Accelerator Compiler Engineer — Inference & Kernels

Ecarx Co.; Greater London; On-site; GBP 100,000 - 150,000; Full time

Job summary

Ecarx Co. in London is seeking an experienced AI/ML systems engineer to design compiler passes for AI accelerators, optimize kernels, and advance LLM inference work. You will bridge AI research and hardware, collaborating with software and hardware teams to push performance and efficiency.

The role requires deep knowledge of C/C++, Python, GPU programming, and DL frameworks, with a track record of optimizing AI workloads on accelerator architectures.

Qualifications

• 5+ years in HPC, compiler dev, or AI/ML systems.
• Strong coding skills in C/C++ and Python.
• Deep experience with GPU/AI accelerator programming (CUDA/OpenCL).
• Hands-on experience with DL frameworks (PyTorch, TensorFlow, JAX).
• Understanding of neural nets and compute patterns (transformers, CNNs, RNNs).
• Experience with performance optimization, profiling AI workloads.

Responsibilities

• Design and implement compiler passes and optimization strategies for AI accelerator code generation (LLVM/MLIR-based).
• Develop high-performance kernel libraries (GEMM, convolution, attention) for our accelerator.
• Lead LLM inference optimization including KV cache management, batching, quantization.
• Collaborate with hardware architects to define ISA, model, and features.
• Integrate compiler stack with PyTorch, ONNX, and other formats.
• Build performance analysis tools and benchmarks to measure model execution.
• Drive optimization across full stack from model to instruction scheduling.
• Contribute to AI software roadmap and accelerator ecosystem.

Skills

C/C++

Python

GPU/accelerator programming

DL frameworks

Transformer architectures

Performance optimization

LLVM/MLIR

Education

BSc/MSc/PhD in CS/EE

Tools

PyTorch

TensorFlow

JAX

Job description

Ecarx Co. in London is seeking an experienced AI/ML systems engineer to design compiler passes for AI accelerators, optimize kernels, and advance LLM inference work. You will bridge AI research and hardware, collaborating with software and hardware teams to push performance and efficiency.

The role requires deep knowledge of C/C++, Python, GPU programming, and DL frameworks, with a track record of optimizing AI workloads on accelerator architectures.