In a previous blog post we introduced the Arm KSA Framework, an evidence-based competency framework detailing knowledge, skills and abilities required for early career software or hardware engineers in the semiconductor industry. In this blog post, we explain how the Framework can help optimize learning, and describe some of its possible applications.
Competency describes an expected level of performance that integrates Knowledge, Skills, and Abilities (KSAs). Educational and training content is a core ingredient to build KSAs, enabling learners to move from one KSA state (with levels for example, introductory, medium, advanced, expert) to another. In a learning context, the former state defines the learning pre-requisites while the latter describes the learning outcomes. We call this journey a learning path.
Figure 1: Learning path illustration.
A learning pathway is a series of KSA ‘hops’ offering scaffolding from a state (with levels) of knowledge/skills/abilities to another, through intermediate KSA states.
Figure 2: Learning pathway illustration.
We can thus define a competency level as a matrix with knowledge (Kx), skills (Sy), and abilities (Az) statements, with relative levels lK…, lS…, lA… as follows:
Figure 3: KSA state matrix representation.
with a learning path or pathways as an operator function that takes a learner from one KSA state, with levels, to another:
Figure 4: Learning path/pathway functional expression
It may seem abstract but this way of thinking about educational experiences becomes increasingly important as we look to optimize learners' knowledge acquisition and skills development. We will revisit this later in this blog post.
The Arm KSA Framework can be mapped against specific software and hardware engineering job roles at different grades/levels e.g., from entry level to senior roles. It can also be mapped against qualification standards (for example, software development technician apprenticeships) or against curricula (for example, bachelor's degree in software engineering). Wider workforce development through analytical research and macro-economic labor policy interventions can also be effectively enabled by KSA frameworks. Moreover, the combination of KSA and generative AI promises to lead to unprecedented levels of accelerated learning. The following sub-sections present these applications in more detail.
The Framework can be used to understand the KSAs applied in software and hardware engineering roles at different grades/levels. This can help enable a common and consistent language in role and job descriptions as well as inform evidence-based talent development efforts within an organization. Indeed, comparing KSAs between an entry level engineering role and a more senior staff engineer role, for instance, can highlight professional development pathways as illustrated in the following figure.
Figure 5: Illustration of KSA role mapping at different grades.
As adopters of the Arm KSA Framework (for example, from industry including the Semiconductor Education Alliance, academia or third sector) adapt and augment the Framework with KSAs that reflect their own organizational requirements or further research evidence, we anticipate this common language/approach will contribute towards a better collective understanding of workforce development requirements across the semiconductor sector. The same principle of course applies to other sectors.
Skills development organizations and agencies all over the world develop qualification standards in a KSA-like approach. The figure below for instance illustrates the standard specification for a digital and technology solutions professional from the UK's Institute for Apprenticeships and Technical Education (IfATE).
Figure 6: Illustration of detail from a qualification standard.
By mapping KSAs from a job role onto qualification standards and vice versa, the gaps and mismatches between qualifications and industry job roles can be bridged in a systemic and sustainable way.
KSA frameworks can also be used to assess the efficacy of education and training programs. At Arm, for instance, we are mapping the Arm KSA Framework against existing and future education content, including Education Kits (courseware), online courses, and books. Gaps in content can be identified and addressed systematically in this manner in an accelerated and agile skill development cycle as illustrated Figure 7.
Figure 7: Accelerated & agile skills provision, circular model as opposed to the traditional pipeline model.
Competency frameworks such as the Arm KSA Framework can be used to validate or invalidate certain hypotheses for example, that social and team skills play a disproportionately higher role in development and promotions at higher grades; or that some assumed competency requirements have adverse Diversity, Equity and Inclusion (DEI) implications. These hypotheses can then inform talent acquisition and development plans.
The frameworks can also be used at a macro level to discover accelerated re/training opportunities from one job category (perhaps disrupted by AI, or where supply is abundant) to another (perhaps a new job category with an acute shortage of supply). Indeed, the quantitative approach to KSA representation suggested in this blog allows for some interesting shortest path/distance analyses and optimizations. The following section provides an example.
The ultimate goal for an education and learning environment is to offer bespoke learning experiences tailored to the needs of each individual. By developing KSA frameworks for various corpuses of competencies with commensurate assessment criteria, we can assess individual competencies at any time and identify/develop the more efficient learning pathways to move an individual from one state of competency to another.
The following figure, for instance, suggests a KSA informed and generative AI powered learning system. In it, a user (or tutor) submits a query, including current and sought KSA statements (including levels). Alternatively, a preceding phase would assess individual competencies against KSA frameworks and provide corresponding KSA statements (with levels) as input to the next phase. A Large Language Model (LLM) trained on KSAs would then suggest optimized learning pathways to move the learner to the desired competency level, which then feeds into a generative AI content generator to produce a user-specific education/training curriculum. The curricula are transformed into vector embeddings stored in an embeddings database, which can then be retrieved for future queries. An assessment schema based on the curriculum suggested serves as a template to assess the efficacy of learning once a curriculum is deployed, the result of which is fed back into the LLM and content generator for fine tuning. Such a virtuous cycle accelerates the discovery and deployment of optimized learning pathways towards successful careers in industry.
Figure 8: A user-tailored accelerated learning environment.
Through this blog post, we explained some of the theory and applications that have driven the development of the Arm KSA Framework . The Framework is underpinning Arm Education’s content development approach as we continue to understand and meet the needs of current and future engineers in the semiconductor industry. We are in discussions with various partners within the Semiconductor Education Alliance and beyond to build, augment, and exploit industry-wide KSAs in various ways including: 1) mapping of KSA frameworks against specific industry job roles at different grades/levels, 2) mapping against qualification standards to iterate on their development, 3) mapping against education and training curricula to bridge the education and skills gaps, 4) wider workforce development, and 5) combining KSA frameworks and generative AI to provide accelerated learning pathways towards successful careers in the semiconductor industry. Reach out to the education team at Arm on education@arm.com if you are interested in collaborating on any of the above.