EduGauge lightning talk at the AECT 2026 Design & Development Showcase
Chicago, Illinois.
Research lab · Since 2019
Research thatstarts with theproblem.
BSHub Research Lab studies how people learn with technology, then builds tools for the problems we find: early-warning analytics, AI learning companions, language coaching for teachers, and more.
Instructors learn a student is struggling only after it’s too late → EduGaugeAI chatbots hand over answers instead of building understanding → Ask JBTeachers must teach in English while still building their own fluency → Dual Learning: Learn TogetherPopular but unsupported education myths spread faster than research → EduFadderCurricula update every few years; industry skill needs change every few months → BridgeConcept maps reveal thinking, but analysing many by hand is slow → jMAPYoung children rarely get playful ways to learn computational thinking → CodeCubsSmall care agencies juggle care, records and compliance on paper → WeCareCourses written in one language exclude learners who think in another → JBT Course TranslatorMany schools still run on paper and scattered spreadsheets → School Management SystemSmall businesses need sales and stock control without enterprise prices → POS SystemInstitutions want Canvas but face licence costs and data-control concerns → Canvas LMS: Self-HostedMoodle is free, but a reliable, well-adopted Moodle isn’t automatic → Moodle: Self-Hosted
The problem ledger
Problems first.
Then the product.
Every product here began as a real problem in a classroom, a campus or a community. Open a row to see who it serves and what it is meant to change. The internal methods stay in the lab.
01Instructors learn a student is struggling only after it’s too lateEduGaugeEarly-warning learning analytics for online courses
The problem
In online courses, struggle is mostly invisible. Instructors usually find out after a missed deadline or a failed exam, when helping is hardest and many students have already disengaged.
Our response
EduGauge works alongside the Canvas LMS. It turns everyday course activity into clear, explainable alerts and next-step recommendations for instructors, students and teaching assistants: what is happening, why it matters, and what to do now.
02AI chatbots hand over answers instead of building understandingAsk JBAn AI learning companion that waits for mastery
The problem
Self-directed learners often move on before they have mastered a concept, and general-purpose chatbots make it easy to get answers without learning anything durable.
Our response
Ask JB is an AI companion for self-directed learning. It checks what you know before and after each topic, scaffolds your thinking, and only moves you on once you reach mastery. It runs on the web, on phones and tablets, and inside Canvas.
03Teachers must teach in English while still building their own fluencyDual Learning: Learn TogetherTeachers and parents grow their English while they teach
The problem
In many rural schools, teachers are expected to teach in English while they are still building their own pronunciation and fluency, and almost none have access to a coach. Parents who want to help with homework face the same barrier.
Our response
A “learn while you teach” platform (Twige Hamwe, “Let’s Learn Together”). Teachers and parents practise pronunciation with instant feedback, using the lessons they already teach. Schools get dashboards, and parents get an interface in their own language.
Who it serves
- Early-grade teachers
- Parents
- School leaders
- Education ministries & NGOs
PilotRead the brief
04Popular but unsupported education myths spread faster than researchEduFadderEvidence at your fingertips when an education myth appears
The problem
Ideas such as “learning styles” spread widely online and shape real teaching decisions, even though research does not support them. Educators rarely have time to check every claim they read.
Our response
A browser extension that notices common educational claims as you read and quietly offers curated, human-reviewed research. It raises awareness without issuing absolute verdicts. Co-created with Dr. Kadir Kozan.
05Curricula update every few years; industry skill needs change every few monthsBridgeAligning what universities teach with what the world needs
The problem
Universities revise programmes on multi-year cycles while the skills employers need keep shifting. Graduates are the first to feel the gap, and institutions lack timely evidence to close it.
Our response
Bridge helps institutions see where their programmes and their students’ readiness match current skills demand, and where they are drifting, so curriculum decisions rest on evidence rather than anecdote.
Who it serves
- University leaders
- Curriculum committees
- Career services
- Students
In developmentRead the brief
06Concept maps reveal thinking, but analysing many by hand is slowjMAPSeeing how learners connect ideas
The problem
Concept maps show how learners connect ideas and reason about cause and effect, but comparing many maps by hand is slow, and closed-ended maps limit what learners can express.
Our response
A web-based causal concept-mapping tool for building, comparing and giving feedback on maps. It has been extended from closed-ended to open-ended maps with AI-supported feedback. Developed with Dr. Allan Jeong.
07Institutions want Canvas but face licence costs and data-control concernsCanvas LMS: Self-HostedYour own Canvas, fully under your control
The problem
Many institutions want Canvas’s teaching capabilities but face licence costs, data-sovereignty requirements, or the need for a sandbox they fully control.
Our response
Design, deployment and administration of independent, self-hosted Canvas LMS instances: servers, security, storage and integrations, plus the training that drives adoption. Running in production at canvas.bshub.us since 2024.
Research areas
Questions we keep asking
From early-warning analytics to computer vision, each area feeds at least one product, and each product sends evidence back into the research.
Learning Analytics & Early Warning
Turning everyday activity in online courses into timely, explainable signals that help instructors act before a student falls behind.
ExploreAI-Generated Adaptive Feedback
When AI feedback helps learning and when it gets in the way, including the timing question: real-time or delayed.
ExploreNLP for Educational Discourse
Automated analysis of online discussions and argumentation, so instructors can see the quality of thinking, not just the volume of posts.
ExploreComputer Vision & Computational Imaging
Image restoration and visual recognition, from despeckling synthetic-aperture radar and ultrasound images to recognising instructional activity in classroom video.
ExploreSelf-Directed & Technology-Mediated Learning
Designing technology that strengthens learners’ own decisions (goals, pacing, mastery) instead of making those decisions for them.
ExploreComputational Mixed Methods & Concept Mapping
Combining causal concept maps, network analysis and qualitative insight to understand how learners connect ideas and think critically.
ExploreInstructional Design & Curriculum Innovation
Systematic, evidence-based design of courses, programmes and multilingual learning experiences, including how AI changes the design process itself.
ExploreAI Policy, Standards & Digital Transformation
Institutional AI policy, international ICT standards, and digital transformation in higher education and government.
ExploreServices
Design, training & advice
The same expertise behind the products is available to universities, schools, governments and organisations, in Rwanda, the United States and anywhere else.
01
Instructional Design
Systematic, evidence-based design of courses, programmes and training, from needs analysis to evaluation, using ADDIE and current learning science.
The challengeCourses are often moved online without being redesigned for how people actually learn there.
02
Training & Capacity Building
Practical, hands-on training and training-of-trainers programmes that leave lasting capacity inside the organisation.
The challengeNew technology arrives faster than staff are trained to use it well.
03
Education & IT Consultancy
Advice grounded in years of running faculties, ICT directorates and university-wide systems, plus doctoral-level research.
The challengeStrategic decisions in education and IT are often made without the evidence or the insider experience to get them right.
04
LMS Infrastructure
Self-hosted Canvas and Moodle, end to end: deployment, security, integration, migration, support and training.
The challengeLearning platforms fail on infrastructure and adoption far more often than on features.
05
Research, Analytics & Evaluation
Rigorous quantitative, qualitative and computational mixed-methods research and evaluation for education and technology programmes.
The challengeProgrammes are launched without a clear way to know whether they work.
Dr. Jean Baptiste MbanzabugaboFounder & Director, BSHub Research Lab
Founder & director
Dr. Jean Baptiste Mbanzabugabo
Most work presents a solution. This lab starts with the problem, and stays with the people who have it.
Learning technologist, computer scientist and higher-education leader working on AI-generated feedback, learning analytics and instructional design
- Degrees
- D.Sc. Computer Science · Ph.D. candidate, FSU
- Leadership
- Dean ×2 · Acting DVC Academic Affairs
- Curricula
- 7 degree programs designed
- Standards
- RSB national technical committee
Selected publications
Evidence, published
- 2026conference
- 2025presentation
Generative AI in higher education: A synthesis of academic leaders’ perspectives
Outstanding Research Paper Award - 2025conference
- 2025journal
- 2025journal
Work with the lab
Have a problem worth solving? Let’s talk.
Research partnerships, pilots, consultancy, training, or a demo of any product.
Latest from the lab
FSU accepts invention disclosure for EduGauge
EduGauge moves toward pilot funding and classroom piloting.
Dual Learning (“Learn English Together”) invention disclosure accepted
Outstanding Research Paper Award, TCC 2025
For a synthesis of academic leaders’ perspectives on generative AI in higher education.