About

Max Fowler is a Teaching Assistant Professor at the Siebel School of Computing and Data Science at the University of Illinois Urbana-Champaign. He holds a Ph.D. in Computer Science from the University of Illinois Urbana-Champaign, earned in 2024. His research areas include Computers and Education. Max Fowler has taught courses such as CS 101 - Intro Computing: Engineering & Science, CS 105 - Intro Computing: Non-Tech, and CS 199 CA2 (CS 199 CA3) - CA Training. He is recognized as an Equity Champion Instructor and an Illini Success Career Influencer, highlighting his contributions to student success and community building within the campus. His work and involvement focus on advancing computing education and fostering an inclusive environment for students in the field.

Research topics

• Computer Science
• Natural Language Processing
• Medical education
• Artificial Intelligence
• Software engineering
• Multimedia
• Mathematics education
• Statistics
• Mathematics
• Engineering
• Medicine
• Linguistics
• Programming language
• Psychology

Selected publications

• The Effect of Transparency on Students' Perceptions of AI Graders

  ArXiv.org · 2026-01-02

  articleOpen access

  The development of effective autograders is key for scaling assessment and feedback. While NLP based autograding systems for open-ended response questions have been found to be beneficial for providing immediate feedback, autograders are not always liked, understood, or trusted by students. Our research tested the effect of transparency on students' attitudes towards autograders. Transparent autograders increased students' perceptions of autograder accuracy and willingness to discuss autograders in survey comments, but did not improve other related attitudes -- such as willingness to be graded by them on a test -- relative to the control without transparency. However, this lack of impact may be due to higher measured student trust towards autograders in this study than in prior work in the field. We briefly discuss possible reasons for this trend.

  PublisherOA PDF

• Using Kotter’s Eight-Stage Model to Understand Departmental Efforts to Broaden Participation in Computing

  ACM Transactions on Computing Education · 2026-03-13

  articleOpen access

  Background : Motivated by the need for a diverse technological workforce, broadening participation in computing (BPC) efforts aim to increase the representation of people who identify as women, African American or Black, Hispanic or Latinx/a/o/e, Native American, Indigenous, persons from economically disadvantaged backgrounds, and persons with disabilities. Research on BPC efforts has highlighted exemplar institutions and activities, but tends to focus on what initiatives computing departments have undertaken and the outcomes of these initiatives. Purpose : Given this prior focus on what initiatives computing departments are undertaking, we propose refocusing on how change happens to increase our collective capacity for impactful change efforts. We apply a well-known organizational change framework, John Kotter’s eight-stage process of leading change, to examine catalysts for change in computing departments, who contributes to this work, and what motivates the work. Doing so can deepen our understanding of BPC efforts and how to enhance them. Theoretical Framework : Kotter’s framework for leading change includes the following eight stages: (1) establishing a sense of urgency, (2) creating the guiding coalition, (3) developing a vision and strategy, (4) communicating the change vision, (5) empowering broad-based action, (6) generating short-term wins, (7) consolidating gains and producing more change, and (8) anchoring new approaches in the culture. Methods : Using a practitioner research approach, we conducted interviews with 13 faculty and staff members across R1 and R2 U.S. institutions via Zoom. Participants were recruited based on their involvement in BPC efforts at their respective institutions. We used inductive and deductive analytic coding approaches to capture how the Kotter framework illuminated participants’ experiences leading BPC efforts. Findings : Kotter’s stages provide a useful breakdown of processes with which to understand and illuminate catalysts for change in computing departments. Across our participants, we see examples of how each stage not only takes shape in different departments, but also how subsequent stages build upon the one(s) before it. Findings reveal a variety of factors that motivate the urgency for computer science (CS) departments to engage in BPC work, along with the importance of top-down leadership and institutional resources. Implications : Kotter’s organizational change model provides an appropriate frame to guide BPC efforts and may be useful to practitioners. Findings from this study illuminate several areas for CS departments to address in order to build capacity for organizational change efforts that support BPC goals. These include examining a variety of departmental data, effectively using meetings and other communications mechanisms, and revising hiring and promotion policies.

  PublisherDOI

• The Effect of Transparency on Students' Perceptions of AI Graders

  arXiv (Cornell University) · 2026-01-02

  preprintOpen access

  The development of effective autograders is key for scaling assessment and feedback. While NLP based autograding systems for open-ended response questions have been found to be beneficial for providing immediate feedback, autograders are not always liked, understood, or trusted by students. Our research tested the effect of transparency on students' attitudes towards autograders. Transparent autograders increased students' perceptions of autograder accuracy and willingness to discuss autograders in survey comments, but did not improve other related attitudes -- such as willingness to be graded by them on a test -- relative to the control without transparency. However, this lack of impact may be due to higher measured student trust towards autograders in this study than in prior work in the field. We briefly discuss possible reasons for this trend.

  PublisherDOI

• A Vision for STEM Higher Education in the Face of AI

  EdArXiv (OSF Preprints) · 2026-05-04

  preprint

  In an extreme challenge to the mission of educational institutions around the world, instructors are concerned --- and in some anecdotal cases, already convinced --- that students are simply using generative AI to produce all their course deliverables, and that no learning is actually taking place. Indeed, research shows that AI is comprehensively capable of completing a variety of science, technology, engineering and mathematics (STEM) courses. Additionally, researchers in multiple fields indicate that AI is able to --- perhaps imperfectly, but fundamentally correctly and immensely quickly --- prove novel results. This development poses a set of critical challenges for the higher education enterprise, as far as to question its future in the face of rapid advancements in artificial intelligence. In this paper, we present a vision on what such a future may look like, addressing four-year degrees, teaching and desired learning outcomes, research and graduate funding, as well as the broader landscape of the future of universities.

  Publisher

• You Don't Need a Data Center to Explain in Plain English! Comparing Open-Source and Propriety LLMs for EiPE Grading

  2026-02-13

  articleOpen accessSenior author

  Code comprehension is an essential ability for Computer Science students, providing a solid foundation for learning programming. An effective approach to evaluating students' proficiency in this skill is through Explain-in-Plain-English (EiPE) problems, which require students to describe the behavior of obfuscated code snippets. Recent advances in large language models (LLMs) have made promising strides toward making autograding EiPE questions feasible. However, prior research has primarily focused on proprietary LLMs, raising concerns over data privacy. In an effort to return autonomy over educational data to instructors and institutions, we investigate the viability of open-source medium-sized language models (MLMs), with parameter counts between 6--100 billion, for EiPE autograding. Our work evaluated several state-of-the-art open-source MLMs on a test set consisting of 620 historical student responses split across 17 EiPE question categories, employing few-shot prompting with three correct and incorrect examples per question. We find several models, such as Llama 3.1 70B Instruct and Qwen 2.5 72B Instruct, that achieve grading accuracy comparable to leading proprietary models like GPT 4o. These results demonstrate that open-source MLMs are promising alternatives for EiPE autograding capable of deployment on local or institution-owned cloud infrastructure. Additionally, we observe that smaller MLMs (less than 32 billion parameters) offer a trade-off between significantly reduced deployment costs and only slightly decreased accuracy, making them well-suited for institutions with limited resources.

  PublisherDOI

• Exploring the Relationship Between Department Characteristics and Computer Science Student Diversity in the US

  2026-02-13

  articleOpen access1st authorCorresponding

  Background: In higher education, efforts to broaden participation in computing (BPC) are typically evaluated at a single institution, which limits our ability to isolate their impacts from those of other factors at the institution.

  PublisherDOI

• Enabling Open Educational Resource Adoption through Integrated Sharing in PrairieLearn

  2026-02-13

  articleOpen access

  This paper introduces the PrairieLearn Question Sharing System (PQSS), which enables instructors to share question generators with other instructors, either as open educational resources or privately. PQSS is integrated into PrairieLearn, an open-source, problem-driven online learning platform. PQSS addresses a critical need for more open-source assessments by making it easier for instructors to share assessments and for instructors to use those assessments. Instructors often do not share questions due to the time it takes to publish them and the lack of recognition for their work. Because it is directly integrated into PrairieLearn, PQSS reduces the aforementioned friction of sharing and using shared questions, and we can report usage statistics to help question authors receive recognition for their work. In this paper, we share design and implementation details of the system, as well as experiences using it to share course content across courses and between universities.

  PublisherDOI

• A Vision for STEM Higher Education in the Face of AI

  2026-05-04

  article

  In an extreme challenge to the mission of educational institutions around the world, instructors are concerned --- and in some anecdotal cases, already convinced --- that students are simply using generative AI to produce all their course deliverables, and that no learning is actually taking place. Indeed, research shows that AI is comprehensively capable of completing a variety of science, technology, engineering and mathematics (STEM) courses. Additionally, researchers in multiple fields indicate that AI is able to --- perhaps imperfectly, but fundamentally correctly and immensely quickly --- prove novel results. This development poses a set of critical challenges for the higher education enterprise, as far as to question its future in the face of rapid advancements in artificial intelligence. In this paper, we present a vision on what such a future may look like, addressing four-year degrees, teaching and desired learning outcomes, research and graduate funding, as well as the broader landscape of the future of universities.

  PublisherDOI

• AI-Supported Grading and Rubric Refinement for Free Response Questions

  2026-02-13

  articleOpen access

  Manually grading free response questions remains a persistent challenge in education. While such questions offer valuable opportunities for student learning and critical thinking, their evaluation often requires substantial time and effort from instructors or teaching assistants. In addition to the grading workload, open-ended responses are susceptible to inconsistencies in scoring and may reflect unclear expectations, both of which can undermine the effectiveness and fairness of the assessment process. To address these challenges, we employed an AI-based grading system integrated in PrairieLearn to automatically evaluate student submissions to free response questions using a predefined set of rubric items. This approach not only streamlines the grading process but also enables direct comparison between AI-generated rubric applications and human judgments, providing insight into alignment and potential discrepancies. These discrepancies provided valuable insight, allowing us to iteratively revise and clarify the rubric items. Our experiences with using the AI grading system across several computing courses suggest that even experienced educators face difficulties articulating rubrics that are both specific and interpretable. We furthermore argue that more attention should be given to the iterative development and evaluation of rubrics.

  PublisherDOI

• Mining Hierarchies with Conviction: Constructing the CS1 Skill Hierarchy with Pairwise Comparisons over Skill Distributions

  2025-02-18 · 1 citations

  article

  Introductory Programming courses teach multiple skills such as 1) explaining the purpose of code, 2) the ability to arrange lines of code in correct sequence, and 3) the ability to trace through the execution of a program, and 4) the ability to write code from scratch. Knowing if a programming skill is a prerequisite to another would assist instructors in organizing their materials such that students encounter and learn new topics using optimal skill sequences. In this study, we used the conviction measure from association rule mining to perform pair-wise comparisons of five skills: Write, Trace, Reverse trace, Sequence, and Explain code. We used the data from four exams with more than 600 participants in each exam from a public university in the United States, where students solved programming assignments of different skills for several programming topics. Our findings matched the previous finding that tracing is a prerequisite for students to learn to write code. But, contradicting the previous claims, our analysis suggested that writing code is a prerequisite skill to explaining code and that sequencing code is not a prerequisite to writing code. Our research can help instructors by systematically arranging the skills students exercise when encountering a new topic.

  PublisherDOI

Frequent coauthors

• Craig Zilles

  University of Illinois Urbana-Champaign

  13 shared

• David H. Smith

  University of Illinois System

  11 shared

• Matthew West

  University of Illinois Urbana-Champaign

  5 shared

• Binglin Chen

  4 shared

• Paul Denny

  Temple University

  3 shared

• Seth Poulsen

  Utah State University

  3 shared

• Kathleen Isenegger

  University of Illinois Urbana-Champaign

  3 shared

• Karrie Karahalios

  2 shared

Labs

• Siebel School of Computing and Data SciencePI

Awards & honors

• Recognition at the Student Success Symposium as an Equity Ch…