2024 Aalto RSE report
In 2024, Aalto RSE has grown compared to 2023, and the work we have done has grown proportionally. We remain busy and in the state of “don’t advertise too much since we would go past our tipping point”.
Summary
Changes from the last report (2023→2024)
We have recruited two more staff members (one internal rearrangement, one new hire) to serve as general-purpose RSEs.
Tracking time spent on every individual project doesn’t seem to work with our current projects. Instead, we’ll try to track general size of projects to understand the distribution of projects among schools, and Halli to understand the money flows.
Our recent growth hasn’t impacted our team cohesiveness yet.
Our current status
More AI-related projects, but by no means exclusively AI. All of our three prongs (programming, data, workflows) can relate to AI.
We maintain a good mix of small, immediate support (SciComp garage) and longer-term projects (Project portfolio)
Even with little advertisement, we are fully booked (but not overwhelmed).
5× time benefit (time we spend vs time academics save), regardless of project size, seems to continue to be true.
Future / What we ask
As always, help in directing to us the academics who most need our services.
In order to serve research needs in other schools, we are looking into an expansion to other schools. The ideal case would be a hire within another school, which would then work with our team but focus on another school.
Current status of Aalto RSE
We have seven dedicated staff, of which three are 100% funded directly by research projects. There are four other Science-IT staff which aren’t paid by RSE funding but dedicate notable amount of time to the service. There are two Science-IT staff who spend a significant amount of time on RSE-related projects.
Types of projects
What do we do? The following figures show the range of activities and roughly where time is spent
One way to classify the types of projects Aalto RSE does, and roughly where our time goes. This isn’t precisely measured but roughly accurate. Real examples given within the cells.
How advanced are our projects? For example, the “basic” level in this figure is things which we’d expect people to have learned in their fields, but still need help with, while “research” is research-level work in projects. We help people at all levels, but ideally would like to focus on the middle two. Ideally, “basic” would be handled more at the academic program level and “practical” could have some more mentoring within research groups, but we realize researchers are very diverse and not everyone can take the same paths in their careers.
There are three main ways our activities are tracked:
RSE projects, bigger activities that are worth adding to our internal project tracker.
Garage support, our daily drop-in time where people can ask any questions. This is often the starting point for larger projects.
The Triton cluster issue tracker provides many projects, usually related to installation or usage of applications.
2024 seems to have had a greater proportion of large projects, taking up a larger amount of time and providing more funding. Our “basic/project funding” RSEs have done more of these large projects, while the dedicated funding staff have had many more smaller AI-related projects.
RSE project stats
A project, for the purposes of this section, is something large enough to keep a record within our issue tracker. Many of our activities are too small for this (see “garage stats” below).
We have realized that with our multitasking and many activities going on (all interspersed and at unpredictable times, due to academic schedules), recording individual time spent on each project doesn’t give good data. Instead, we will track the overall size distribution of projects and use the finance system to track time spent for funding matters (which includes our largest projects). EU projects also require a separate time tracking system.
Number of RSE projects 2020-2024: 260
Number of new RSE projects, 2024: 71
Distinct contact people 2024: 69 supervisors, 77 contacts
Number of distinct schools: 5 (all except CHEM)
Projects per year. The data is interesting, but since it tracks “RSE project issues created” and the standard for what should be an issue changes over time, it shouldn’t be relied on too much.
We see that the general trend for number of contacts is increasing. This data isn’t completely accurate, though.
We see that the School of Science continues to be the primary customer base, as is expected since it’s the only one providing basic funding.
We also see that most time is going to the School of Science, followed by activities that support all Aalto in general, and then the School of Electrical Engineering. Because time is not tracked rigorously, this should only be used as a relative measure.
“Tasks” are labels for what projects are about - projects can have multiple. This shows roughly the kind of things we do.
Like above, but with the time distribution. Again, the absolute values are not accurate but this shows the relative distribution.
We see that we are shifting more towards larger projects. This is a real effect, but also “don’t bother recording small and garage projects” is a big part of it, too.
A word cloud of the one-sentence summaries of projects.
Garage stats
Our Scientific computing garage is a daily session where any researcher at Aalto can drop by and ask us for help. This is where we meet the most people and has taken over as the “small project” system.
13 distinct departments logged + all six schools. (We haven’t recorded full data on everyone’s department.).
Every school represented.
542 recorded visits.
Not every visit is recorded, the total visits is estimated around twice as many.
Departments/units of garage customers.
Schools of garage customers.
Academic positions of garage customers.
Word cloud of garage summaries.
Triton issue stats
We don’t have stats here, but many of the RSE-related ones are related to installation and use of various software.
Funding status
Our funding has been good: In summary, we had enough paid project work that we have needed to expand our team, since we were at the point of not being able to accept new paid projects. The general principle remains “There should be less than 2FTE of staff time paid by basic SCI funding”.
IT Services continues to fund work that comes from outside the School of Science. There is one new major project that is expected to provide at least 1FTE/year of projects over the next years. We have worked out funding practicalities for EU projects.
The statistics below are a measure of worktime with vacations and sick leaves distributed equally across the other sources. 1FTE is one person working for one year.
Summary of funding flow of the RSE service, measured it FTEs of funding. We had 5FTE at the start of the year and 7FTE at the end, so the total FTE was 5.5 throughout the year. Overall, basic funding from SCI was 20%, or 1.1 FTE. This is slightly increased because our new hires can’t begin their new projects immediately. “Dedicated” are staff who are paid 100% by one major funding instrument and work for many related projects.
Entire year: 3 FTE paid directly by a research project, not listed below.
End of September: (before new hires began) 0.6 FTE basic funding, .98 FTE from project funding (.51 FTE from projects directly, .48 FTE from IT Services).
End of year: 1.1FTE basic funding, 1.4 FTE from project funding (of which .84FTE from projects directly and .58FTE from IT Services)
Future plans
The increase in AI-related research will probably continue to give us more work in the future, even among our existing research community.
Our two new hires this year will do significant work for a major new funding instrument, so a lot of their time is already planned to be funded in the next year.
We are working to bring other schools onboard, so that they can hire their own research software engineers to serve their communities. An example of how this can work can be found in RSE work rotations.
As always, we have a good number of customers from our existing work community, but there are so many more potential users who don’t know of us. The academics who don’t currently know about us are probably the ones that need us most.