Download Full Issue PDF
Introduction to Volume 17, Issue 2
David Joinerpp. 1–1
A brief introduction to this issue of the Journal of Computational Science Education from the editor.
Volume 17 Issue 2 — November 2026
Contents
The Computational and Data Science Exchange (CDSExchange): A Model for Building Faculty Capacity in Computational and Data Science Education
Katharine Cahill, Karen Rios-Soto, Tandabany Dinadayalane, Rachel Vincent-Finley, Linda Akli, Ahlam Tannouri, Zhuoying Wang, and Lisa Garbrechtpp. 2–5
https://doi.org/10.22369/issn.2153-4136/17/2/1BibTeX
@article{jocse-17-2-1,
author={Katharine Cahill and Karen Rios-Soto and Tandabany Dinadayalane and Rachel Vincent-Finley and Linda Akli and Ahlam Tannouri and Zhuoying Wang and Lisa Garbrecht},
title={The Computational and Data Science Exchange (CDSExchange): A Model for Building Faculty Capacity in Computational and Data Science Education},
journal={The Journal of Computational Science Education},
year=2026,
month=nov,
volume=17,
issue=2,
pages={2--5},
doi={https://doi.org/10.22369/issn.2153-4136/17/2/1}
}
Copied to clipboard!
Small and lesser-resourced institutions often face challenges launching or expanding undergraduate Computational and Data Science (CDS) course offerings. The CDS-Exchange addresses this challenge by building a faculty community of practice across institutions. Core components of the project include: webinar series, regional workshops, a fellows program, and annual symposia. These diverse activities result in broad outcomes, such as multi-institution participation, strong repeat engagement, and positive self-reported gains in understanding and teaching relevance.
Adapting for Adoption: Tailoring HPC Training for Resource-Constrained Communities
Bryan Johnston, James Carson, Charlie Dey, Jennifer M. Schopf , Christopher J. Whalen, and Grace Kebirungipp. 6–9
https://doi.org/10.22369/issn.2153-4136/17/2/2BibTeX
@article{jocse-17-2-2,
author={Bryan Johnston and James Carson and Charlie Dey and Jennifer M. Schopf and Christopher J. Whalen and Grace Kebirungi},
title={Adapting for Adoption: Tailoring HPC Training for Resource-Constrained Communities},
journal={The Journal of Computational Science Education},
year=2026,
month=nov,
volume=17,
issue=2,
pages={6--9},
doi={https://doi.org/10.22369/issn.2153-4136/17/2/2}
}
Copied to clipboard!
Successful High-Performance Computing (HPC) adoption in resource constrained environments requires more than hardware deployment; it requires training models that are explicitly designed for local realities rather than transplanted from well-resourced contexts. This position paper argues that leadership-class HPC curricula must be re-engineered into pragmatic, context-aware training pathways if emerging HPC communities -- particularly in Africa -- are to achieve sustainable operational capability. Drawing on the adaptation of the Texas Advanced Computing Center's (TACC) leadership-class training into an onboarding program for African HPC sites, we illustrate how global best practices can be reframed into actionable modules aligned with infrastructural, staffing, and operational constraints. We contend that such contextualisation is not optional but foundational for long-term HPC sustainability, and we outline principles for designing training that empowers institutions to deploy and operate HPC systems sustainably and effectively within resource-limited environments.
Introducing HPC-Enabled Computational Research Training in Africa: A Pilot Study
James Afful, Titus Nyarko Nde, Onyeka Onyenemezu, Brian Kyanjo , Benjamin Atribawuni Asaaga, and Joseph Oppong Akowuahpp. 10–19
https://doi.org/10.22369/issn.2153-4136/17/2/3BibTeX
@article{jocse-17-2-3,
author={James Afful and Titus Nyarko Nde and Onyeka Onyenemezu and Brian Kyanjo and Benjamin Atribawuni Asaaga and Joseph Oppong Akowuah},
title={Introducing HPC-Enabled Computational Research Training in Africa: A Pilot Study},
journal={The Journal of Computational Science Education},
year=2026,
month=nov,
volume=17,
issue=2,
pages={10--19},
doi={https://doi.org/10.22369/issn.2153-4136/17/2/3}
}
Copied to clipboard!
Access to computational training and research-computing mentorship remains uneven across global scientific communities. This paper reports a pilot study of the Center for Computational Research and Education (CCRE), a training initiative designed to introduce students and early-career researchers in Africa to computational research through a staged model of open recruitment, structured preparation, mentored project work, and reflective evaluation. The pilot was advertised through LinkedIn, WhatsApp, and Instagram for more than two months, and 24 applicants registered. Twelve applicants were selected for a mandatory four-module Google Classroom preparatory course; six completed all required modules and advanced to the mentored group-project phase. Evaluation used pre- and post-program survey summaries from the final cohort. Baseline results showed strong motivation to build practical high-performance-computing (HPC) skills, connect with a research-computing community, learn collaborative software practices, and explore computational approaches in participants' domains. They also showed substantial structural barriers: Most respondents reported no prior HPC access and no or limited formal training in HPC, parallel computing, or research computing. Post-program respondents gave an average overall rating of 4.20 out of 5, and all five indicated that they would definitely recommend the program. Respondents nevertheless continued to report barriers related to HPC access, local mentorship, and computing cost. The CCRE pilot is presented as an educational-practice model whose principal contribution is a documented and critically evaluated pathway for introducing computational research training under constrained infrastructure and mentorship conditions.
Expanding the National AI Research Resource: Building Capacity and Community through Sustainable Research Pathways
Mary Ann Leung, Connie L. McNeely, and Danielle Jacquespp. 20–25
https://doi.org/10.22369/issn.2153-4136/17/2/4BibTeX
@article{jocse-17-2-4,
author={Mary Ann Leung and Connie L. McNeely and Danielle Jacques},
title={Expanding the National AI Research Resource: Building Capacity and Community through Sustainable Research Pathways},
journal={The Journal of Computational Science Education},
year=2026,
month=nov,
volume=17,
issue=2,
pages={20--25},
doi={https://doi.org/10.22369/issn.2153-4136/17/2/4}
}
Copied to clipboard!
Focused on facilitating artificial intelligence (AI) research through the U.S. National Science Foundation (NSF) National AI Research Resource (NAIRR), the Sustainable Research Pathways (SRP) program provides an effective model for building an innovative, interactive, and wide-ranging community of research-competent computing sciences professionals. Aimed at expanding access to NAIRR for higher education faculty and students, SRP provides networking opportunities and cultivates collaborations with related researchers using structured, evidence-based approaches built on over 10 years of experience. More specifically, SRP incorporates three principal programmatic and interactive training strategies: 1) broadly recruit students and faculty-student teams from a wide range of backgrounds and institutions, 2) establish active collaborative research relationships through an iterative and successively more intensive matching process, and 3) catalyze the development of relevant competencies and of systemic and participatory research capacities. Accordingly, SRP operates to pave the way for further expansion of NAIRR. SRP is aimed at developing skilled talent poised to address the pressing challenges facing the AI workforce. It has the potential to develop into a new and effective paradigm for expanding access to NAIRR while making significant progress in AI accessibility and providing a broader spectrum of perspectives on the development of AI models and methods. The program builds strong connections among the existing NAIRR researchers, institutions, and individual participants across the U.S., especially those who would not typically engage in NSF-sponsored AI research. SRP expands NAIRR by democratizing, increasing the variety of AI talent, and advancing safe, secure, and trustworthy AI. The program explores changes to the approaches and practices of AI research through collaborations between NAIRR projects and new and emerging researchers—-a crucial activity for future AI efforts.
T3-CIDERS Train-the-Trainers: Strengthening CI Skills for Cybersecurity Research and Education through Iterative Improvements in Second Cohort
Wirawan Purwanto, Mohan Yang, Peng Jiang, Dorothy X. Parry, Belle Li, Kristin Herman, Nolan Lovett, Shanan Chappell Moots, Masha Sosonkina, and Hongyi Wupp. 26–33
https://doi.org/10.22369/issn.2153-4136/17/2/5BibTeX
@article{jocse-17-2-5,
author={Wirawan Purwanto and Mohan Yang and Peng Jiang and Dorothy X. Parry and Belle Li and Kristin Herman and Nolan Lovett and Shanan Chappell Moots and Masha Sosonkina and Hongyi Wu},
title={T3-CIDERS Train-the-Trainers: Strengthening CI Skills for Cybersecurity Research and Education through Iterative Improvements in Second Cohort},
journal={The Journal of Computational Science Education},
year=2026,
month=nov,
volume=17,
issue=2,
pages={26--33},
doi={https://doi.org/10.22369/issn.2153-4136/17/2/5}
}
Copied to clipboard!
T3-CIDERS is a multi-year train-the-trainer program that develops university faculty, researchers, and students as future trainers that are equipped to integrate advanced cyberinfrastructure skills, including high-performance computing, machine learning, big data analytics, and parallel programming, into cyber-security research and education. This paper describes the iterative redesign of T3-CIDERS from its first cohort (2024) to its second cohort (2025–2026), motivated by systematic evidence collected through pre/post surveys, daily reflections, observations, and interviews. Challenges identified in Cohort 1, such as unequal participant preparedness, participant fatigue from a dense schedule, limited interactivity, and insufficient community building drove a set of evidence-based program improvements. For Cohort 2, the pre-training phase was substantially expanded to cover all six technical modules and a foundational pedagogy module. The in-person institute was restructured for more substantial interactivity and engagement, adopting a rhythm of mini-lectures, hands-on CI learning, guided discussions, and studio-based lesson development. Community-building activities were strengthened through social events during the institute, structured mentorship meetings, and a new community webinar series. Early findings from Cohort 2, drawing from reflections and direct feedback from the participants, show increased participation in the program beyond the institute, as well as confidence and eagerness to teach CI to others.
In Search of a Better Search: Using RAG on Cornell Virtual Workshop Content
Zilu Wang, Susan Mehringer, and Christopher J. Cameronpp. 34–38
https://doi.org/10.22369/issn.2153-4136/17/2/6BibTeX
@article{jocse-17-2-6,
author={Zilu Wang and Susan Mehringer and Christopher J. Cameron},
title={In Search of a Better Search: Using RAG on Cornell Virtual Workshop Content},
journal={The Journal of Computational Science Education},
year=2026,
month=nov,
volume=17,
issue=2,
pages={34--38},
doi={https://doi.org/10.22369/issn.2153-4136/17/2/6}
}
Copied to clipboard!
The Cornell Virtual Workshop (CVW) is a continually updated HPC training platform with about 40 roadmaps and more than 2000 pages, organized in a three-level hierarchy of roadmaps, topics, and pages. The CVW's traditional keyword search works poorly for users who lack domain terminology or want navigational guidance rather than a specific page. We describe a RAG-based chat-bot that routes each query to a retriever matched to the appropriate level of the content hierarchy, ensuring that navigational queries return information about CVW structure and factual queries return specific page content. The chatbot is built with LangGraph and uses a hybrid BM25 and FAISS retrieval pipeline with cross-encoder reranking. A fixed-path pipeline provides prompt-injection defense, query rewriting, post-generation URL verification, and graceful handling of out-of-scope queries.
A Documentation Catalog for the AI Ecosystem: Design of the ICICLE AI Training Catalog
Amit Vyas, Carlos Guzman, Neelima Savardekar, Mary Thomas, Michelle Miller, Alfonzo Morales, Joe Stubbs, Christian Garcia, and Beth Plalepp. 39–45
https://doi.org/10.22369/issn.2153-4136/17/2/7BibTeX
@article{jocse-17-2-7,
author={Amit Vyas and Carlos Guzman and Neelima Savardekar and Mary Thomas and Michelle Miller and Alfonzo Morales and Joe Stubbs and Christian Garcia and Beth Plale},
title={A Documentation Catalog for the AI Ecosystem: Design of the ICICLE AI Training Catalog},
journal={The Journal of Computational Science Education},
year=2026,
month=nov,
volume=17,
issue=2,
pages={39--45},
doi={https://doi.org/10.22369/issn.2153-4136/17/2/7}
}
Copied to clipboard!
As artificial intelligence becomes increasingly important across scientific disciplines, there is a growing need for scalable platforms that support training, knowledge sharing, and workforce development. The ICICLE AI Institute, funded by the National Science Foundation, focuses on democratizing artificial intelligence and cyber infrastructure through open education, community engagement, and shared software infrastructure. This paper presents the design and architecture of the ICICLE Training Catalog and demonstrates how the platform supports broader community engagement and workforce development through practical use cases and user personas. The catalog aggregates educational materials, tutorials, and technical documentation from multiple institutions into a centralized and structured repository. We describe the website design, documentation design principles, and early usage evidence from privacy-preserving Umami analytics. We also discuss how the Training Catalog supports workforce participation development and how it can be leveraged by different user groups, including students, researchers, educators, and research software engineers. Finally, we highlight lessons learned from developing a collaborative training platform across institutions and discuss future directions for expanding its impact.
Understanding Computational Science and Engineering (CSE) and Domain Science Skills Development in National Laboratory Postgraduate Internships
Morgan Fong, Hilary Egen, Marc Day, Kristin Potter, and Michael J. Martinpp. 46–56
https://doi.org/10.22369/issn.2153-4136/17/2/8BibTeX
@article{jocse-17-2-8,
author={Morgan Fong and Hilary Egen and Marc Day and Kristin Potter and Michael J. Martin},
title={Understanding Computational Science and Engineering (CSE) and Domain Science Skills Development in National Laboratory Postgraduate Internships},
journal={The Journal of Computational Science Education},
year=2026,
month=nov,
volume=17,
issue=2,
pages={46--56},
doi={https://doi.org/10.22369/issn.2153-4136/17/2/8}
}
Copied to clipboard!
Background and Context: Harnessing advanced computing for scientific discovery and technological innovation demands scientists and engineers well-versed in both domain science and computational science and engineering (CSE). However, many universities lack the resources needed to provide integrated training. National laboratories offer internship opportunities capable of developing these skills. We present an evaluation of federally-funded post-graduate internship outcomes in CSE at a national laboratory. We seek to answer three questions: 1) What computational, research, and professional skills do students improve through internships at the selected national laboratory? 2) Do students gain knowledge in domain science topics through their internships? 3) Do students' career interests change after these internships.
Leveraging HPC in University Courses: A Cross-Disciplinary Workforce Development Model
F. Alex Feltus, Jon C. Calhoun, Rong Ge, Jill Gemmill, Carlos Toxtli Hernandez, Bing Li, Todd Allen Price, Justin Talbot, and Minjae Woopp. 57–64
https://doi.org/10.22369/issn.2153-4136/17/2/9BibTeX
@article{jocse-17-2-9,
author={F. Alex Feltus and Jon C. Calhoun and Rong Ge and Jill Gemmill and Carlos Toxtli Hernandez and Bing Li and Todd Allen Price and Justin Talbot and Minjae Woo},
title={Leveraging HPC in University Courses: A Cross-Disciplinary Workforce Development Model},
journal={The Journal of Computational Science Education},
year=2026,
month=nov,
volume=17,
issue=2,
pages={57--64},
doi={https://doi.org/10.22369/issn.2153-4136/17/2/9}
}
Copied to clipboard!
This paper presents a cross-disciplinary workforce development model in which nine faculty members at Clemson University integrated the Palmetto2 supercomputer into seven courses spanning Bioinformatics and Computational Biology, Business Data Analytics, Computational Chemistry, Computer Science and Parallel Architectures, Automotive Engineering and AI Applications, Healthcare Data Science, and HPC Systems Hardware. A democratized access policy—removing departmental and financial barriers to cluster accounts-enabled broad adoption across student populations ranging from wet-lab biologists to working business professionals. Common pedagogical challenges included HPC onboarding, software dependency management, and large-dataset transfer; shared solutions included priority course scheduling, scaffolded desktop-to-cluster exercises, and modular training materials developed by Clemson's Research Computing and Data Engagement (RCDE) team. Discipline-specific computational profiles varied from large-dataset I/O in bioinformatics and healthcare to GPU-intensive deep learning in automotive AI and intensive floating-point computation in chemistry and parallel architectures. Three transferable principles emerge: democratized access, guaranteed priority scheduling for courses, and domain-contextualized scaffolded exercises. Students completing these courses gain Linux proficiency, job scheduling, parallel programming, and largescale data analysis skills directly applicable to research and industry careers in an economy increasingly dependent on dataintensive computation.
Low-Cost Embedded Computer Vision and Neural Network-Enabled Autonomous Navigation for a Four-Wheeled Mobile Robot Using an ESP32 and Arduino Framework
Abhiram Kaakarlapp. 65–72
https://doi.org/10.22369/issn.2153-4136/17/2/10BibTeX
@article{jocse-17-2-10,
author={Abhiram Kaakarla},
title={Low-Cost Embedded Computer Vision and Neural Network-Enabled Autonomous Navigation for a Four-Wheeled Mobile Robot Using an ESP32 and Arduino Framework},
journal={The Journal of Computational Science Education},
year=2026,
month=nov,
volume=17,
issue=2,
pages={65--72},
doi={https://doi.org/10.22369/issn.2153-4136/17/2/10}
}
Copied to clipboard!
This paper presents the design, construction, and testing of a low-cost autonomous navigation system built on the Freenove 4WD robot car. An ESP32-CAM microcontroller handles image capture and onboard machine learning while an Arduino Uno controls the four DC motors, with all hardware costing under 103 US dollars. A convolutional neural network trained through Edge Impulse was compressed to an 18.2 KB quantized model that runs directly on the ESP32, classifying 32 by 32 pixel grayscale camera frames into five sign categories: stop sign, left turn, right turn, T intersection, and forward arrow. An OpenCV image processing pipeline running alongside the neural network performs Canny edge detection and Hough line transform for lane tracking and pixel-density-based obstacle avoidance. Navigation decisions are encoded into short serial packets with CRC error checking and sent from the ESP32 to the Arduino, which converts them into PWM motor signals. Across 500 controlled test runs, the robot correctly identified signs 91.78 percent of the time under even lighting and 69.46 percent under uneven lighting. The full decision loop from camera capture to motor response averaged 66.6 milliseconds. A systematic failure analysis across all 500 trials categorized errors into hardware, environmental, and software causes; five targeted fixes reduced the overall failure rate from 36.8 percent to 11.4 percent. The project also served as a comprehensive self-directed computational science education exercise, and the paper includes a dedicated student reflection documenting technical lessons, methodological insights, and educational impact. All firmware, training data, and experimental records are publicly available for reproduction and improvement by other students and educators.