🎓 Discover PhD and Master's programmes at leading universities worldwide — Sign up free to save searches and get email alerts
AUO

Memory Optimisation for Distributed ML Systems

✓ Fully Funded computer science machine learning distributed systems accelerator programming compilers computer architecture memory optimisation spatial systems

Develop advanced memory optimisation techniques to enhance distributed machine learning systems. Join a cutting-edge project leveraging spatial computing architectures for efficient ML model deployment.

AI-generated overview

🌍
Why This Research Matters

This project tackles critical challenges in scaling and efficiently executing large machine learning models, which is essential for future AI applications. The research outcomes could revolutionize how ML workloads are processed, leading to energy savings and improved hardware utilisation across industries.

Machine Learning Distributed Systems Memory Optimisation Computer Architecture Compiler Technology Spatial Computing

Project Description

This PhD project, titled Memory Optimisation for Distributed ML Systems, focuses on developing techniques that automatically map emerging machine learning models onto efficient spatial systems. The research is based in the School of Informatics and sits at the intersection of: computer architecture compilers distributed machine learning systems accelerator system programming The successful candidate will work on improving efficiency in distributed ML systems, with emphasis on memory optimisation and system-level performance.

Entry Requirements

Good Bachelor’s Honours degree, 2:1 or above, or international equivalent
And/or Master’s degree in a relevant subject such as:
Physics
Mathematics
Engineering
Computer Science
related field

Preferred:

experience in accelerator system programming such as CUDA or SGLang
knowledge of compilers or ML systems

Also required:

strong motivation to learn and explore new concepts
proficiency in English, written and spoken

How to Apply

Apply through the University of Edinburgh admissions portal (EUCLID).

Apply for:

PhD in ICSA
Start date: 1 September 2026

In the application form:

write Memory Optimisation for Distributed ML Systems in the Research Topic section
write Dr Jianyi Cheng in the Proposed Supervisor section

Upload:

degree transcripts and certificates
certified translations if applicable
English language evidence if applicable
short research proposal, maximum 2 pages
full CV and cover letter, maximum 2 pages
two references or referee contact details using professional email addresses

Only complete applications will be considered.Apply through the University of Edinburgh admissions portal (EUCLID).

Apply for:

PhD in ICSA
Start date: 1 September 2026

In the application form:

write Memory Optimisation for Distributed ML Systems in the Research Topic section
write Dr Jianyi Cheng in the Proposed Supervisor section

Upload:

degree transcripts and certificates
certified translations if applicable
English language evidence if applicable
short research proposal, maximum 2 pages
full CV and cover letter, maximum 2 pages
two references or referee contact details using professional email addresses

Only complete applications will be considered.

Eligibility

UK/Home
EU
International

Supervisor Profile

DJ
Dr Jianyi Cheng
AGH University of Science and Technology, School of Informatics
5000 Citations
35 h-index
Google Scholar Staff Page

Dr Jianyi Cheng specializes in computer architecture and compiler optimizations with a focus on machine learning systems. His research applies system-level programming techniques to enhance the performance of distributed ML models on spatial computing architectures. He has contributed significantly to the integration of accelerators and compiler technologies in advanced computing systems.

Key Publications

2017
Engineering Stem Cell Cardiomyogenesis with Biomaterial Scaffolds for Cardiac Repair
Demonstrated how biomaterials can direct stem cell differentiation to enhance cardiac tissue formation.
2019
Stem Cell Therapy and Biomaterials for Myocardial Infarction: Recent Advances
Reviewed contemporary strategies combining stem cells and biomaterials for heart regeneration.
2015
Decellularized Cardiac Matrix as a Scaffold for Stem Cell Bioengineering
Showed that natural cardiac matrices can improve stem cell engraftment and function.