2026 ASEE Annual Conference & Exposition

AI as an Intern: A Curriculum Framework for Supervising Vibe Coding in Data Science Education

Presented at DSAI-Session 1: Data Science Program Design and Curriculum Frameworks

The emergence of what is increasingly referred to as ‘vibe coding,’ formalized here as an AI-assisted development framework grounded in natural language prompting, iterative reasoning, and contextual evaluation, may represent an important shift in Data Science (DS) education. Instead of emphasizing syntax mastery and procedural execution, vibe coding reframes computational work as an interactive modeling process between human reasoning and intelligent systems. This paradigm can help rebalance the implementation of algorithms, analytical modeling, and theoretical understanding in DS instruction. This article develops a curriculum framework that integrates vibe coding into Data Science programs to support conceptual learning, model interpretability, and quantitative rigor. Automating routine programming tasks may allow faculty to reallocate instructional time toward the mathematical and statistical foundations of the discipline, including probability theory, linear algebra, optimization, and statistical inference. These areas form the computational core of Data Science and underpin reproducibility, verification, and uncertainty quantification. The curriculum structure is organized around three technical components. Conceptual alignment ensures that AI-assisted competencies are mapped to DS learning outcomes in data literacy, algorithmic reasoning, and model transparency. Pedagogical implementation embeds AI-supported programming exercises into courses such as Applied Data Modeling and AI Collaboration in Data Science, emphasizing prompt precision, data validation, and computational reproducibility. Ethical and analytical assurance introduces evaluation protocols for AI-generated code using statistical benchmarking, bias detection, and verification metrics. Integrating vibe coding into DS curricula can enhance accessibility, reduce redundant technical overhead, and strengthen students' ability to connect computational processes with statistical reasoning. This approach offers an instructional model that aligns automation with analytical depth, helping scientists prepare data to design, interpret, and validate AI-driven systems with both technical and mathematical precision. The approach may also be adapted, with discipline-specific modifications, to Computer Science curricula.

Authors

Dr. Irene Tsapara http://orcid.org/https://0009-0006-1819-2639 National University [biography]

Dr. Irene Tsapara is an academic leader, researcher, and educator specializing in Data Science, Artificial Intelligence, and computational learning theory. She serves as Academic Program Director for the Doctorate in Data Science at National University, where she oversees curriculum design, research supervision, and program accreditation. Dr. Tsapara holds a Ph.D. in Mathematics with specialization in Computational Learning Theory, Machine Learning, and Artificial Intelligence from the University of Illinois.
Her current research focuses on the integration of AI-assisted learning frameworks, including vibe coding, into Data Science and interdisciplinary education. She develops scalable curricular models that align AI literacy, statistical reasoning, and ethical governance with foundational mathematical rigor. Her work emphasizes the use of large language models (LLMs) in promoting conceptual understanding, transparency, and reproducibility in computational practice.

Dr. Tsapara has over two decades of teaching and programming experience and has contributed to the development of graduate and doctoral curricula across data science, machine learning, and applied analytics. She also collaborates with the National AI Research Institute (TILOS) and several academic-industry partnerships focusing on responsible AI, data governance, and interdisciplinary education.
Her professional mission is to reimagine Data Science education as a connective discipline, one that unites computational precision, mathematical depth, and human-centered AI ethics to prepare graduates for leadership in an intelligent, data-driven world.
Danae Nikki Vassiliadis University of Chicago [biography]

Danae Vassiliadis is a data science practitioner and educator whose work focuses on applied machine learning, data systems, and AI-driven decision-making, with an emphasis on human behavior and interaction with intelligent systems. She serves as a Teaching Assistant in the Master of Science in Applied Data Science program at the University of Chicago, where she supports graduate-level instruction in big data systems, cloud computing, and machine learning.

She is also a Data Science Consultant at Accenture, where she leads generative AI learning and upskilling initiatives and contributes to the development of proprietary GenAI tooling for enterprise applications. Her work examines how individuals and organizations interact with AI systems, including the design of frameworks for the evaluation and deployment of large language model–based solutions, with attention to user behavior, trust, and decision-making.

Danae holds a Master of Science in Applied Data Science from the University of Chicago and dual bachelor’s degrees in Industrial Engineering and Psychology from Northwestern University. Her interdisciplinary training informs her approach to the design of AI systems that account for behavioral variability.

Her current research focuses on AI-mediated programming and education, including human–AI collaboration and the integration of large language models into data science training. She is particularly interested in the effects of generative AI on learning behavior, authorship, and problem-solving, as well as in the development of evaluation frameworks that support rigor and transparency. More broadly, her work examines how data science education and practice can incorporate AI while maintaining methodological and analytical foundations.
Andrew D’Amico Northwestern University [biography]

Andrew D'Amico is an entrepreneur, AI systems architect, and applied data science practitioner whose work lies at the intersection of artificial intelligence, systems design, analytics, and engineering education. He is a consultant with Northwestern Analytics Partners, where he contributes to AI and analytics initiatives spanning intelligent decision support, quantitative modeling, and agent-based systems.

He holds a Master of Science in Data Science in Artificial Intelligence from Northwestern University and undergraduate degrees in Philosophy and English Literature from Loyola University Chicago, with a concentration in Philosophy of Science. He serves as a teaching assistant in Northwestern's Data Science program, supporting instruction in Business Process Analytics, Decision Analytics and Operations Research, Data Engineering, and Quantitative Finance in Data Science. He is a certified Project Management Professional (PMP) and PMI Disciplined Agile Senior Scrum Master (DASSM), with delivery experience across Scrum, Waterfall-Predictive, and Disciplined Agile methodologies.

His current research focuses on two convergent areas. The first is AI-mediated software development, including vibe coding and collaborative programming frameworks that are reshaping how programming is learned, guided, and evaluated. The second is assessment and pedagogy in the age of AI, with particular attention to authorship, process-based evaluation, educational validity, and the preservation of rigorous learning outcomes in environments shaped by generative systems. More broadly, his work examines how institutions can integrate AI in ways that are methodologically sound, educationally credible, and responsive to the changing structure of technical and intellectual labor.

Drawing on both technical and humanistic training, he brings a multidisciplinary perspective to the governance and institutional adoption of AI in engineering and data science education.

Note

The full paper will be available to logged in and registered conference attendees once the conference starts on June 21, 2026, and to all visitors after the conference ends on June 24, 2026

« View session

For those interested in:

Faculty