SURF and researcher working together

Open Call: High Performance AI4Science Projects

SURF invites researchers from Dutch research institutions to apply for consultancy and computational resources for the development of a scientific use case in machine learning-assisted data analysis, modelling, and interpretation.

The call at a glance

Open to researchers from all disciplines.

Receive consultancy, technical support, and hardware infrastructure from SURF, in collaboration with Intel experts, for state-of-the-art machine learning techniques applied to scientific data. Our contributions include: 

  • support from experts on scientific machine learning.
  • support from experts on parallelization, GPU programming, performance optimization, etc.
  • computational hours on the latest Intel Data Center CPUs and GPUs, and/or the Dutch national supercomputer Snellius.
  • optional publicity of research findings promoted by SURF and/or Intel.  

We welcome independent and skilled researchers at all stages of their careers to apply.

Project duration: 6-12 months. 


- 30 April 2024, 23:59 CET: proposal deadline.

- 31 May 2024: selection deadline for winning proposals.

- 1 June - 1 September 2024: preferred start of the project.

Submit an application

Purpose of this call

Our call for proposals is centred around the integration and application of Artificial Intelligence (AI) in scientific workflows. Our primary goal is to encourage and facilitate the adoption of AI technologies, especially in projects where AI has the potential to significantly advance the field. By providing our support, we aim to overcome initial obstacles that can slow down or even prevent the development of such projects. 

One key aspect of our mission is to maximise hardware efficiency, thereby amplifying the power of AI and optimising the effectiveness of existing infrastructure. Beyond these objectives, our broader aspiration is to cultivate an innovation-driven community, one that contributes not only to scientific progress but also to the development of scientific outputs that are reusable, reproducible, and open (subject to privacy constraints). 

This call is an invitation to researchers who share our vision of integrating AI into the fabric of scientific discovery, pushing the boundaries of research through collaborations and cutting-edge methodologies.

What we offer

The successful applicants will gain computational hours tailored for their project needs. The specific allocation will be determined on a case-by-case basis, taking into account the nature of the problem, the data involved, and the chosen approach. 

SURF and Intel are able to provide expert guidance in various areas of AI, programming, and high-performance computing (HPC), including but not limited to time series modelling, generative architectures, scaling models to the billion-parameter level, training, fine-tuning, and deploying foundation models, few-shot learning approaches, multi-modal models, and fast inference. 

Moreover, we provide support for distributed training across multiple GPUs, including the partition of model states and CPU offloading. Research findings may benefit from enhanced visibility, courtesy of promotion by SURF and Intel. In the case of Master's or Bachelor's students interested in pursuing an internship with SURF, we offer a remuneration that is competitive with current market rates.

What we look for

The project calls for a well-defined use case where data analysis, modelling, and interpretation plays a crucial role in addressing a scientific problem. Applicants are required to provide a sufficiently large scientific dataset appropriate to their proposed use case. The dataset should be publicly accessible and free from any personal or sensitive information. The availability of the proposed dataset will be a key factor in the selection process for successful proposals, as it will significantly influence the feasibility and potential impact of the project. 

Our objective is to evaluate an array of machine learning methodologies to accurately capture and model the behaviour and statistical properties inherent in the dataset. This involves not only identifying patterns for classification, regression, and anomaly detection, but also generative modelling (e.g. Bayesian inference and data augmentation), as well as symbolic regression for discovering underlying equations or relationships in datasets.

Who can apply

Proposals are welcome from researchers, PhD candidates, Master’s and Bachelor’s students who are affiliated with or pursuing a degree in Dutch academic institutions, research centres, or other public knowledge-driven organisations. 

An overview of most eligible organisations can be found here. In case of uncertainty regarding eligibility, please reach out to us for further clarification.


For questions about this call, please contact Yue Zhao, high performance machine learning advisor at SURF, via

Submit an application