Skills, Careers and Training

Strategic Rationale and Evidence Base

Strategic Importance

HEP and RA research has become computing intensive, at least since three decades. Large infrastructures and large investments are needed to collect, store and analyse the datasets produced. This requires dedicated teams of experts with specific skills and capabilities to design and operate the required computing systems and software stacks; failing that, the expected impact is a reduced physics exploitation capability for the planned billion Euro investments. In many cases critical pieces of software depend on individuals or small groups of people often in non-permanent positions. Together with workforce ageing, migration to industry and lack of recognition for research software engineering work, this poses a risk to research.

Training, supporting and retaining computing experts is of critical importance for current and next generation initiatives, even more so in a scenario where the same capabilities are in great demand from industry. In several cases researchers are lacking training in research software engineering and will receive the necessary training only when joining a research collaboration. Community established training in this respect is even more vital.

Current State

Computing experts (either by formation or trained inside the initiatives) suffer from inadequate career paths. They are seldom offered permanent positions, and those that do rarely have a strong career path. Academic positions in physics are not open to them, and in the best case they are retained as “technical staff”, typically on “soft money” (renewed year on year without long term job security). Computing experts also face an attractiveness deficit relative to industry (which offers higher salaries and more stable positions) and structural under-recognition in scientific outputs (rarely first author on the resulting publications, often not credited as authors at all). The work of research software engineers or researchers who code also often lacks recognition of their contributions within the communities which hampers career progression in addition.

Community Evidence

See D6.1 Section 2.5 (Consolidated Gaps and Requirements), Gap #13 (Workforce sustainability and development): the problem of “Workforce, funding, and sustainability” is listed as one of the most critical gaps; the highlighted solution consists of the “... development of a coordinated European funding/governance models, in alignment with national contributions, and in the investment in training and career pathways for research software engineers and data stewards”. The Physics Briefing Book “Input for the 2026 update of the European Strategy for Particle Physics”, Section 12.6, explains the problem from the angle of High Energy Physics. In the conclusions, the problem of career paths is detailed as “... in the long term, this is more worrying than technological trends, and finding solutions should be a field priority”.

Technical Foundation

The problem is not strictly technical, but rooted in the historical way in which computing related tasks have been executed in the target physics domains. Up to 3 decades ago, computing was a post-facto solution, designed after data taking and long after the initial design. This was possible due to its (much) reduced complexity with respect to other technical issues (e.g. building and operating a detector). Today, the size and complexity of scientific computing has increased to a level of criticality comparable with other technical aspects, with the overwhelming majority of experts acknowledging this. Nevertheless, research and academic institutions have failed to keep pace and continue to undervalue scientific computing experts, treating them as replaceable. The result is a fast turn-around of less experienced computing collaborators, who leave after contributing critical software to their communities. Worse still is when this happens at a later career stage, where computing contributions are much more significant. Entire systems need to be replaced, without the knowledge and expertise of the original expert. This issue directly affects the sustainability of large scientific initiatives, and their capability to produce scientific results. It also hampers European strategic efforts to find synergies across scientific computing domains, as the few experts that do exist are too busy mitigating these issues. Finally, while it’s desirable to have some early career individuals leave academia for industry, for those who leave later in their careers the transition is typically much less efficient for industry and more stressful on the individual.

European Alignment

The career problem for science is now recognized in European initiatives, including the EuroHPC JU and the EOSC, and is leading to promoting positions as technical specialists and data stewards. Still, at the academic level, the career paths offered to computing experts (either by formation or by specialization) are less attractive than, for example, in the construction of equipment and in data analysis. Sovereignty of computing initiatives and skills is an important aspect which needs to be preserved, in a global scenario where major non-European industries drive the technology innovation. The EOSC EVERSE project, as one example, strives to establish best practices in research software engineering underpinned with services to provide a coherent and connected infrastructure for training materials and events as well as providing recognition for research software engineers and trainers.

Target Outcomes and Impact Assessment

Strategic Outcomes

Inform and reiterate among research and academic institutions the need and importance of research personnel fluent in large scale computing; prepare and fund appropriate career paths for such experts and provide means for the recognition of research software engineering work. Feedback to universities and engineering schools on training needs in scientific collaborations

Success Indicators

Lower total costs for the computing aspects of European scientific initiatives, driven largely by the ability to create longer-term, more sustainable computing plans, both during the design and operations phases. Improved knowledge sharing in computing across scientific domains thanks to the existence of a small cohort of career computing experts providing a stable and strategic vision for European scientific computing. Of course this cannot be realised by SPECTRUM alone, but must come from a domain-level realization of the issue and its possible solutions. With well-defined career paths, smoother transitions between academia and industry mean individuals destined for industry will start their careers earlier.

European Impact

A solid cohort of domain scientists with extended computing skills, able to confront the data deluge expected from next generation scientific initiatives. In perspective, since not all of them will be able to remain in the public research domain, a cohort of scientists is moving to the private / productive sector, and participating in its innovation.

Implementation Pathway and Stakeholder Roles

Implementation Phases

• Short-term actions (deliverable under current initiatives: RI Train, HPC Train, EuroHPC Centres of Excellence) raise awareness of academic and research partners to the problem of providing good career paths to computing experts. Medium-term actions (deliverable through new project proposals or integration into university curricula) target structural change in education and recognition.
• Long term actions require a reshaping of the funding and personnel policies in academic and research institutions.

Stakeholder Roles

Scientific Communities

• Responsibilities: clarify their present / future needs in computing, and the needs in expert personpower; clarify the risks attached to their vacancy. Communicate the challenges to be solved, and the possibility to directly influence scientific results.
• Benefits: secure long term computing positions, for the benefit of decades-long initiatives. Increase the potentially interested community.

Research Infrastructures

• Responsibilities: ensure sufficient staffing and smooth operations (for research and beyond).
• Benefits: long-term planning of resources allows to participate in decades-long initiatives, and to have personnel with a deep knowledge in the systems and thus able to react / plan in a more effective way.

Computing & Data Service Providers / e-Infrastructures

• Responsibilities: Provide comprehensive training on the efficient use of heterogeneous resources, covering hardware acceleration (e.g. GPUs), software best practices (e.g. containers, portability), and sustainable operations; offer secondment opportunities for Research Software Engineers to embed within operational centres. Facilitate contacts between the scientific user communities and partners from industry, in order to show further career possibilities.
• Benefits: Improved infrastructure efficiency through optimised user workflows; reduced support burden; and a user community capable of exploiting advanced platform features.

Policy Makers / Funding Bodies

• Responsibilities: fund computing research positions at academia and research institutes. Ensure career paths for researchers designing / operating computing systems.
• Benefits: a more predictable landscape in computing allows for a long-range planning of initiatives. A solid scientific computing infrastructure in specific scientific domains allows for easier support for the long tail of research. The realization of a cohort of experts which can partially fuel innovation in the productive system.

Resource Needs

• More funds dedicated to computing experts (by formation, training or inclination) in academic and research institutes. Long-term plans guaranteeing sustained support.

Critical Dependencies

All the other priorities, but especially highlighted for

• P07 AI
• P08 Software Portability
• P09 Open Science and Reproducibility
• P10 Software Preservation and Sustainability
• P11 Community Collaboration and Co-Design

-