- The paper introduces an interdisciplinary ADAPT model that transforms DS/AI education with modular, project-based approaches across diverse academic units.
- It details a flexible framework replacing traditional courses with 1-credit, skills-based modules and backward design, ensuring real-world applicability.
- The study demonstrates robust scalability and high satisfaction metrics, highlighting innovative institutional strategies for workforce preparedness.
The ADAPT Model: An Institutional Mechanism for Interdisciplinary and Workforce-Relevant Data Science and AI Education
Introduction
The proliferation of data science (DS) and AI as critical, workforce-relevant domains has generated significant curricular challenges in higher education. Conventionally, institutional responses have been siloed, with efforts restricted to designated departments or isolated professional programs. The All-campus Data Science and AI Project-based Teaching and Learning (ADAPT) model, developed by the Data Science and AI Academy (DSA) at North Carolina State University, offers a novel university-wide, interdisciplinary framework specifically targeting scalable, flexible, and workforce-aligned instruction in DS and AI (2604.02597). This essay synthesizes the core features, implementation strategies, and implications of the ADAPT model with a focus on its impact, scalability, and future directions for DS/AI education.
Institutional Innovation: The Academy as a Transversal Node
Unlike traditional academic units (departments, schools, or institutes), NC State's Academy model was created as an agile, boundary-crossing entity to synergize existing strengths and fill critical gaps in DS/AI education, research, and outreach. Operating in a "skunkworks" mode, the DSA bridges 12 colleges and nearly 90 departments, unlocking access to instructional resources, research enablement, and cross-campus programming for every academic constituency, including undergraduates, graduates, non-degree seekers, and K-12 learners.
A principal innovation is the replacement of rigid, credit-intensive courses with modular, 1-credit DS/AI courses at three distinct levels (200, 400, 500), coded by prerequisite skills rather than formal courses. This design dramatically lowers the threshold for interdisciplinary engagement and allows for both vertical and lateral stackability within major/minor requirements, certificates, and elective spaces. Integration of instructors from academia, industry, and government (42%, 19%, and 12% in Spring 2026, respectively) ensures continuous alignment with workforce trends and authentic real-world applications.
The ADAPT Model: Structure and Pedagogical Underpinnings
The ADAPT model is conceptually anchored in three pillars:
- Project-Based Learning (PBL)
- Ten Common Learning Elements
- Workforce Preparedness
Project-Based Learning
All DSA courses are structured around substantive projects, employing formative and summative assessment strategies that emphasize performance, knowledge application, and iterative feedback. The methodology replaces conventional assessments with assignments mirroring authentic professional workflows (scoped projects, presentations, code review, reflective writing), thus mapping educational experiences onto actual DS/AI practice. Backward design techniques are systematically utilized within instructor professional learning communities to scaffold learning outcomes toward project deliverables.
Ten Common Learning Elements
Derived from national recommendations (NASEM, ASA), these elements are grouped into Data Perspectives (data as information; professional identity; exposure to diverse careers), Data Practices (acquisition, curation, validation, design, ethics), and Data Discoveries (frontier questions; impactful outcomes). Instructors exercise autonomy in implementation, but all are required to explicitly embed these elements into course design. Iterative, community-driven revision of these elements provides dynamic responsiveness to stakeholder feedback and evolving field trends.
Workforce Preparedness
This component operationalizes the development of DS/AI agency and professional identity. Students select data contexts and analytic methods relevant to their interests and future careers, optimizing both motivation and skill acquisition. The design encourages exploration of consequential phenomena (e.g., broadband access, algorithmic bias), critical tool selection (programming languages, visualization, modeling), and communication practices essential for workforce integration. Peer-led Course Collaboration Leaders (CCLs) extend mentoring and support, further decentralizing instructional scaffolding.
Implementation Metrics and Institutional Impact
Since its 2021 inception, the DSA curriculum expanded from 5 to 29 sections per semester with enrollment consistently meeting section caps (average 25 students/section) by Spring 2026. DSA course integration into over 188 unique majors across all NC State colleges demonstrates robust cross-disciplinary penetration. Distinctive features include:
- Adoption of 1-credit courses as major electives, minor/certificate components, or stand-alone credentials
- Rapid expansion of offerings (e.g., Introduction to R/Python, AI Ethics, Network Analysis, Predictive Analytics, Sports Analytics, Algorithmic Fairness)
- Creation of new, interdisciplinary minors and certificates (e.g., Data Science in Business, Engineering Analytics, K-12 Education)
- Articulation of DSA courses within existing credential structures by numerous departments, broadening both undergraduate and graduate engagement
- Institutional partnerships for non-credit, workforce-centric programming (e.g., NC Department of Health and Human Services, community colleges)
Quantitative Evidence and Professional Development Outcomes
The ADAPT model has achieved high satisfaction metrics in pilot workshops disseminated to community college instructors (overall satisfaction: 4.61/5), and participants reported high confidence for curricular integration of ADAPT principles (avg: 4.49/5). NSF-funded research efforts (awards #DUE-2313644 and #DGE-2222148) have supported empirical evaluation and iterated improvement of program efficacy, professional learning communities, and collaborative curricular development.
Theoretical and Practical Implications
The ADAPT model's scalable, modular, and interdisciplinary approach provides an operational blueprint for addressing the persistent challenge of democratizing DS/AI education. Its design offers two major theoretical implications:
- Flexibility and Customization: The stackable, skills-based course structure accommodates heterogeneous learner backgrounds and interests, aligning with contemporary models of adaptive, personalized learning in STEM.
- Authentic Professionalization: PBL, combined with workforce-centered project choice and assessment, generates verifiable transfer of skills and professional identity construction—both essential for robust labor market outcomes.
On a practical level, the Academy model foregrounds the necessity of institutional infrastructure that is sufficiently agile to manage, support, and sustain cross-campus curricular innovations. This is especially relevant as data science and AI increasingly permeate non-STEM fields, requiring durable interdisciplinary partnerships and alignment with rapidly shifting professional standards.
Prospects for Future Development
Given its positive trajectory, ADAPT's key elements are likely to disseminate across peer institutions, driven by:
- Increased demand for DS/AI competencies in non-traditional sectors (humanities, education, healthcare, policy)
- Expansion of PBL methodologies into graduate-level and professional continuing education, including custom corporate training
- Further integration with K-12 education, utilizing the existing ADAPT-aligned CS/DS curricular materials
- Ongoing empirical research into best practices in interdisciplinary, project-based STEM instruction and DS/AI identity development
Scaling challenges will arise around maintenance of instructional quality, assessment rigor, and instructor professionalization as course demand grows. Rigorous outcome tracing (career pathways, skill retention, workforce integration) will be critical for evaluation and sustainable expansion.
Conclusion
The ADAPT model, as implemented by NC State’s DSA, establishes a robust, evidence-driven paradigm for interdisciplinary, workforce-relevant DS/AI education. Its core contributions include modular, project-intensive curricula; explicit, research-based learning objectives; and decentralized pathways for instructional and professional innovation. As both a mechanism and a transferable model, ADAPT provides a compelling framework for institutional transformation in response to the exigencies of widespread data and AI literacy demands (2604.02597).
