Researcher SoC architecture for Machine Learning
What you will do
- You will explore the power, performance and area (PPA) of dedicate Machine Learning architectures to benchmark them in imec’s Machine Learning technology offering.
- Teaming up with various experts inside imec, you will explore system and architectural trade-offs for new technologies across the whole system, from new (synapse or neuron) devices up to novel compute in memory technology up to 3D integration.
- You will assess which technology enablers will bring which gains in the overall system, for different AI workloads - ranging from standard CNNs to cutting edge reinforcement learning or online learning.
- You will maintain and build links with imec’s commercial and academic partners and follow-up joint research projects with them
- You will help drive imec’s technology and architecture roadmap to build the semiconductor technology for AI of the near and far away future.
What we do for you
Who you are
- We are looking for a PhD in Electronics or Computer Architectures, or a master with at least 3 years relevant industrial experience.
- We value a background of digital implementation and SoC architecture for one or more application domains (Mobile, Server, …) and experience with specifications for different parts of SoC in different market spaces.
- You are familiar with deep learning and machine learning algorithms and can communicate with our algorithm experts.
- You have worked in multi-disciplinary teams, ideally both interacting with architects, hardware designers as well as algorithm and software teams
- You have experience with system exploration simulator software (gem5, mcpat, Platform Architect) or are willing to learn to use these tools.
- Since you will work in an international environment, you have a very good knowledge of English.