Consultancy

Pictures of NeSI research software engineers. From top to bottom: Chris Scott, Alex Pletzer, Wolfgang Hayek, Maxime Rio.
Attribution: 
NeSI’s Consultancy Service provides scientific and HPC-focused computational and data science support to research projects across a range of domains.

Need support with your research project?

If you would like learn more about NeSI's Consultancy service and how you can work with NeSI Research Software Engineers on a project, please contact us (support@nesi.org.nz) to set up an initial meeting. We can discuss your needs and complete a Consultancy application form together.

Researchers from NeSI collaborator institutions (University of Auckland, NIWA, University of Otago and Manaaki Whenua - Landcare Research) and those with Merit projects can usually access consultancy at no cost to themselves, based on their institution's or MBIE's investment into NeSI.

What do we do?

Examples of outcomes we could assist with (this list is general and non-exhaustive):

  • Workflow parallelisation – allowing more inputs to be processed simultaneously, usually by different instances of the software running side by side
  • Software parallelisation – use of technologies such as OpenMP or MPI to process one single input more quickly by one instance of the software
  • Code optimisation – redesign of algorithms to improve overall speed or efficiency of resource (CPU or memory) use
  • Improving I/O performance – work to speed up reading from or writing to the disk, or to reduce the amount of data that must be read or written
  • Porting to GPU – accelerate code by offloading computations to a coprocessor
  • Creating/Improving a visualisation workflow - specialised visualisation solutions or developing interactive web-based dashboards
  • Translating Python/R/Matlab code to C/C++/Fortran for faster execution
  • Improving software sustainability – introducing best practices such as version control and unit testing
  • Conducting an Exploratory Data Analysis - use visualisations and statistical methods to summarise the main characteristics of a dataset in light of scientific questions
  • Designing and fitting explanatory models - adapt or create statistical models to estimate latent variables distributions from noisy observations and expert knowledge
  • Designing and fitting predictive models - train machine learning models to predict variables of interest on unlabeled datasets
  • Cleaning data - assist with automating the identification and correction of erroneous values in a dataset

What can you expect from us?

During a consultancy project we aim to provide:

  • Expertise and advice
  • An agreed timeline to develop or improve a solution (typical projects are of the order of 1 day per week for up to 4 months but this is determined on a case-by-case basis)
  • Training, knowledge transfer and/or capability development
  • A summary document outlining what has been achieved during the project
  • A case study published on our website after the project has been completed, to showcase the work you are doing on NeSI

What is expected of you?

Consultancy projects are intended to be a collaboration and thus some input is required on your part. You should be willing to:

  • Contribute to a case study upon successful completion of the consultancy project
  • Complete a short survey to help us measure the impact of our service
  • Attend regular meetings (usually via video conference)
  • Invest time to answer questions, provide code and data as necessary and make changes to your workflow if needed
  • Acknowledge NeSI in article and code publications that we have contributed to, which could include co-authorship if our contribution is deemed worthy
  • Accept full ownership/maintenance of the work after the project completes (NeSI's involvement in the project is limited to the agreed timeline)

Previous projects

Listed below are some examples of previous projects we have contributed to: