AI Performance Engineer, Life Sciences & GPU Optimization

Advanced Micro Devices; Helsinki; On-site; EUR 90,000 - 130,000; Full time

Benefits offered by this job

AMD benefits at a glance

Job summary

Advanced Micro Devices (AMD) seeks an AI Performance Engineer, Life Sciences to contribute to foundation models and agentic AI systems for drug discovery. You will work with a multidisciplinary team of ML researchers, computational chemists, and software engineers to deliver production‑grade AI solutions.

This role offers hands‑on work in deep learning performance optimization, cross‑domain collaboration, and the ability to translate requirements into scalable software that can have real-world

Qualifications

• Master’s or PhD in a related field with practical experience in ML or HPC.
• Experience delivering ML or simulation solutions and pipelines.
• Strong understanding of deep learning architectures and transformers.

Responsibilities

• Optimize ML models through profiling and bottleneck analysis across GPU architectures.
• Analyze models for compute and memory requirements and optimize for training/inference.
• Perform kernel, framework, and hardware level optimizations.
• Work with deep learning, transformers, and physics-based simulation models.
• Profile workloads on current hardware and improve efficiency and scalability.

Skills

Python proficiency

Deep learning knowledge

English communication

Education

Master’s degree or PhD in CS/CE/CompSci/Applied Math

Tools

PyTorch

CUDA

TensorRT

HIP

Triton

OpenCL

Nsight

Job description

Advanced Micro Devices (AMD) seeks an AI Performance Engineer, Life Sciences to contribute to foundation models and agentic AI systems for drug discovery. You will work with a multidisciplinary team of ML researchers, computational chemists, and software engineers to deliver production‑grade AI solutions.

This role offers hands‑on work in deep learning performance optimization, cross‑domain collaboration, and the ability to translate requirements into scalable software that can have real-world