By SSI Fellow Kim Martin.
Last year, I was gifted the opportunity to attend the Society of Research Software Engineering's Sixth Annual Conference (RSECon22), held in Newcastle (UK) in September 2022; my travel costs were covered by a combination of a conference travel bursary along with an additional 'RSE Worldwide' grant from the Society, and my conference attendance was covered by the Software Sustainability Institute (SSI). I was attending the conference to present a talk on my ongoing efforts to establish a sustainable RSE Group at Stellenbosch University in South Africa, as well as to join a panel discussion on 'RSE Worldwide'. I was and still am very grateful for the opportunity to attend: I was able to meet members of RSE Groups from across the UK (many of whom I'd connected with online during the previous year); I was able to ask questions and have discussions with RSE Group leaders on how their Groups were founded and are run; I could attend multiple excellent presentations in person (and socialise with the speakers and audience afterwards). The experience was heady and reinvigorated my commitment to the goal of building an RSE support service at my home university.
When I learned about the Research Software Engineer (RSE) role for the first time in 2021 (from a keynote given by Dr Heidi Seibold at the useR! 2021 online conference, the term 'RSE' was almost unknown in Africa, despite past efforts by Talarify in collaboration with the SSI; the RSSE Africa network (founded by Peter van Heusden) was dormant at the time, although it has recently been reinvigorated with additional support from Talarify. According to a Research Software Alliance (ReSA) representative I communicated with in March 2022 (who was involved in a survey of RSE-related activities in the Global South, no RSE Group of the type that is now common in the UK existed at any African University.
My PhD is in biomedical science (tissue engineering), although at the time I had recently made the switch to a research group studying wood formation in eucalypt trees (another kind of tissue development!), mainly because I was looking for ways to move away from 'wet lab' work and into computational work. Learning about the RSE role was a bit of a revelation, and gave me a clear sense of the direction my career could take (staying in a research environment, using the knowledge gained from my long years of scientific and laboratory training, and yet developing the computational and software engineering skills that I had found myself drawn towards). Attending the excellent online September RSE21 conference introduced me to the wonderful RSE community, and strengthened my conviction that Research Software Engineering was both a meaningful and satisfying career and also a critically-important part of the academic ecosystem.
Chicken or egg?
My first goal after attending the online 2021 RSE conference was to connect with any existing local community of RSEs, as I felt that my development as an RSE would be best done in the context of a local group of peers who I could learn from and potentially collaborate with. While my attempts at networking within my South African university quickly made it clear that the term 'RSE' was effectively unknown here (although I did of course find 'proto-RSEs' like myself), I also found myself talking with many researchers from different Faculties who - on hearing me explain RSE - were quick to insist that they needed the support of RSEs in their research.
With the evidence of need (and support from members of my university community) mounting, my goal shifted to finding a way to provide RSE support to researchers at Stellenbosch University. From what I had learned of the way that UK RSE Groups operated, it seemed that there was a chicken-or-egg problem in establishing an RSE Group de novo: for a group to exist, it requires a source of funds to support RSE salaries... but without an existing culture of writing RSE support into academic grant applications, where do the funds come from? Similarly, an RSE group needs to be able to recruit computationally-skilled and research-experienced people to work as RSEs... but without awareness of RSE as a valid and exciting potential career option, where do the people come from?
Priming the pump
I decided to approach the problem as a 'pump-priming exercise': if I could find a way to give my university a taste of the kind of services and support RSE Groups typically provide, that could be a way of escaping the chicken-or-egg problem and leapfrogging towards the self-sustaining models that have been successful at universities in the UK and elsewhere.
As I understood it, RSE Groups typically provide three main types of support: Capacity development (training researchers in coding and related skills); Consultation (assisting researchers in identifying how RSE support could enable them to do more and better research); and Coding support (writing software and improving existing codebases). My first target was to deliver training under the umbrella of RSE@SUN, and to that end, I turned to the excellent Carpentries organisation for support. I was awarded Carpentries Certified Instructor status in October 2021, and swiftly set out to capture details of which Carpentries training materials were in demand in my university (Python and R, unsurprisingly, although I insist on adding Git on principle), and started running workshops with the help of collaborators. To my delight, RSE@SUN was gifted a Sponsored Silver Carpentries Membership this year, which enables me to train additional instructors and provide more workshops.
Providing scalable coding support to researchers is more challenging under this bootstrapping strategy, but I have been fortunate to have cultivated a strong network of support, including a number of far-seeing Professors in fields ranging from computer science and informatics to data science and bioinformatics, along with staff within the Research IT service at Stellenbosch University. I've been exploring what different models of support provision could look like, including more distributed and dynamic approaches. Earlier this year, one high-profile Professor advertised for 'RSE Postdocs', including my name on the advert, and offered to allow me to use some of his new recruits' time in providing RSE support within the university. The Society of RSE has graciously included one of this Professor's postdocs in their RSE mentoring scheme, thus helping establish the RSE identity here.
I submitted my SSI Fellowship application at the end of October 2021, and (wonderfully), in January 2022 I was named an SSI Fellow - one of the first international (non-UK) SSI Fellowships in the 10-year history of the program. In service of my pump-priming plan, my Fellowship idea is to use my funds to provide consulting to Stellenbosch researchers from UK-based RSEs. Through RSEs advising local researchers on how RSEs could assist them in their work, and by providing the researchers with assistance to write RSEs into their upcoming grant applications, I believe that this could be a way of establishing the prerequisites for a sustainable local RSE service. Similarly, the consultations would provide lists of skills that each researcher would need to look for in their RSE partner(s) going forward, making it possible to identify local 'proto-RSEs' who have the skills but (so far) lack the sense of identity as an RSE... thus helping build the local RSE community.
With the increased credibility lent by the SSI Fellowship, I presented an informational pitch to Stellenbosch University's Deputy Vice-Chancellor (DVC) Research and Innovation about the need for an RSE Support Service at Stellenbosch University. My pitch was well-received, and with support from the office of the DVC, I had the opportunity to present at one of the meetings of the Deans and Vice-Deans of Research from multiple Faculties in May 2022.
Encouraged by the public support I was receiving, including substantial support from Dr Jan Greyling of the Stellenbosch AgroInformatics Initiative, I launched the rse.sun.ac.za website in June 2022 and presented a talk titled 'Research Software Engineering (RSE) as a component of a healthy academic ecosystem' at a National Institute for Theoretical and Computational Sciences (NITheCS) Colloquium - seeing it as a good opportunity to raise the profile of RSE Nationally.
The opportunity to attend RSECon22 in person was very special, particularly given how significant it had been to attend the online September RSE21 the previous year. My talk was titled 'Bootstrapping an RSE Group at a university in South Africa', and the support from the audience was great: including offers of practical support in providing training and other services to my growing community. The chance to connect in person with members of the UK RSE community was wonderful, and talking with people gave me a real boost in motivation and - crucially - ideas. Being awarded the ‘RSE Rising Star Award’ was a cherry on the cake… and also a valuable symbol of the support I have from the RSE community, which in itself helps me in my mission by providing me with additional credibility in South Africa.
One of the ideas that was born at the conference, and which is coming to fruition now: I will be doing an 'RSE Roadtrip' in the UK in May and June (funded by the SSI), visiting a number of university-based RSE Groups to compare and contrast their histories, strategies, and operational realities. The first draft of the interview guide listing the type of topics and questions I'm aiming to include is published on Zenodo, along with a link to the Google Doc which is open for comments (which I intend to take into account, listing contributors on the next version of the document). I hope to publish my observations in a form which will be of use to the RSE community (including the nascent African community). I also intend to take my learnings into account in developing a proposal for RSE@SUN going forward, for which I am fortunate to have the backing of the DVC Research and Innovation at Stellenbosch University.
Want to discuss this post with us? Send us an email or contact us on Twitter @SoftwareSaved.