# Aalto RSE info chat

* Attendees from Aalto, TU-Delft,

## About the session

The Aalto RSE service has gone on well for about a year now. Interested in hearing inside news about how it works internally, how you might do something similar, or how we can work together? Join us at for an informal discussion. https://scicomp.aalto.fi/rse/ (see also "Video: Aalto RSE status report, May 2021" in the 'internal documents' section of that page).

Richard Darst is an RSE founder of the Aalto Scientific computing RSE team, this short meeting would be mostly about sharing Aalto's approach to support services around scientific computing.

### What was the motivation behind the intent of putting such a team together? Could you explain what is the philosophy behind the design of your services? Any learnings and experiences to share?

- Richard
- HPC binded together, hardware and infrastructure
- We wanted to make it more accessible to people. Lets get more people to use the tool. A motivation was to make the technology more accessible to people that need to do computational work. Example: Made someone independent after some support, the time saving is the actual benefit. So the support of an expert makes a difference.
- There are two scenarios: do it with them and for them.
- Diversity argument: do we want to bias good work towards those who happen to have acquired the right skills years ago?
- If the University take care of diversity then we should enable people with different backgrounds to take advantage of the resources.
- Thomas:
- Automating things and helping with few lines of code can save them a lot of time, this points to workflows.
- Jarno
- Most of the projects are software development, if you understand how science goes and modelling, you can go really far. We rely on low hanging fruits. The customer or researcher is the domain expert and we are the expert in software development, so we are supposed to translate domain expertise needs into technical requirements.
- Keeping talented people in academia. RSE name and role didnt exist. This position creates a career path for many people that drop out because they spend so much time learning software development that they didnt have the same ammount of papers and citations.
- Provide help in using infrastructure and helping with the technical side, then more people can do computational related work.

### How are your services structured and why have you structured in such a way?
- Richard
- "do it for them", "do it with them and they take over", "fix it up after someone else has done it".
- Team is not organized at a university level, we report to the dean not ICT services. We report to researchers and work with researchers. We are not that centralized, we are in diferent buildings and spread.
- The ideal case would be that we provide some general support, there would be local RSEs or people with expertise in software engineering. If other faculties could also recruit and create local RSEs that would be ideal.

- Jose
- Then working with the critical mass of advanced users and researchers that could be potential RSEs or take that role is a promising model.

- Maurits
- What projects do you prioritize? How do you prioritize projects
- The setup is different, while you started bottom-up we have started top-down which makes things challenging. Things can get dilluted at a University level, the enthusiasm can be easily get dilluted.

- Richard
- The structure is very related to the funding, balance of prioritizing strategic projects and supporting many people. Right now we are funded by the departments/school, so bias is towards short-term projects for everyone. Long-term projects should be funded by the projects themselves, which in theory would let us expand more. We serve the schools that fund us, not the entire university.
- The benefit is to help people for all the schools/facluties.

- Jarno:
- We create a space where everyone can join that connects different stakeholders. Building a network that doesnt exist is the most difficult. But the garage/(daily RSEs open office hour) really helps to stay connected.
- Things can evolve from a simple consultation, to a short term project(5 to 6 working days), to a long project, to helping people in hiring an RSE and training them.

- Thomas:
- Support in garage consultations, can evolve from small consultations to projects.
- We have the RSE hour, everyday where people come with questions, its completely open, people can come and that makes the barrier lower. Half an hour of consultation support a little bit of code, you can convert consultations into a project. More than half of the projects come from the garage.

- Jarno on HPC training:
- We started running workshops as a way of support researchers and project where they didnt have expertise to be able to use the cluster. We were running 2 to 4 software carpentries.

- RD:
- When you have infrastructure and hardware people needs to use it. Things have to be integrated, hands on support, training and access to hardware.
- - picture here of three-way feedback loop, https://scicomp.aalto.fi/about/

Overview of Aalto's services structure that combines infrastructure, training and hands on-support


### Could you introduce a bit your team and what they do at the University? What is the team composition and type of RSE profiles?

- We are 2-4 people depending on how you count us.
- We have funding mainly from the School of Science, but also now from university level.
- PhDs, hired from similar fields as we serve. Basically former computational scientists.
- How many are you? How many do you think the University really needs?
- 2 RSEs, two part-time RSEs. Broader Aalto Scientific Computing team has a total of 8-10 (including RSEs).

## About expanding Scientific computing knowledge transfer and communities
- We have tried to have "poweruser groups" to support experts within groups, who thus go and support the rest of their groups. So far, despite trying, we haven't had time to do this well.
- What challenges have you faced during your growth as a team?
- RD: I would say outreach to a broad audience, since good communication takes time and it is hard when you are bottom-up. Even though we have no lack of projects.

### What is the relationship between this team, the coderefineries initiative and the RSE-nordic community?
- We found good international projects to support, and contributed to them because we knew it would only help us. We have quite a bit of presence there, but they aren't "our" projects.
- Nordic RSE group works for RSEs and is not really made by "us" (Aalto), but of course the people are the same. CodeRefinery is also a separate project, but we find it useful for supporting researchers.

### With regard to the coderefineries specifically how do you think the TU Delft team could contribute to them?

- Thomas: Its about building it with researchers, they have to see the benefit of it, it has to do with the individual groups faculties, and have a very low barrier to get your initial consultations and support. Having university funding to offer basic support and then illuminate researchers about he costs of having such RSE role, help them on a basic level, show at least with anecdotal examples of what they could do.

- RD: Coderefineries is for people that already pogram but need to use the best programming tools at the next level (better version control, testing, documentation, modularity, etc.)