EdTech Archives EdTech Archives Proceedings of the Learning Engineering Research Network Convening (LERN 2026)

Open Repository for AI Models as Learning Engineering Components

Jim Goodell

Abstract

HuggingFace.co is one of the most influential platforms for developers building, sharing and using machine learning models used by thousands of organizations, however, finding models applicable to learning is nearly impossible and the interface is highly technical. LearningFace.ai is offered as an alternative that making AI models more accessible to learning engineering teams. As the “hugging face for learning analytics” it offers a bridge between research and practice. It empowers researchers to get their work published and give learning engineers access to those discoveries for data-driven decision-making to optimize learning. The beta open access site is piloting a new approach to democratizing AI learner state detection as “learning engineering components”, packaged in a form that can be implemented in various learning systems. This poster will give a demonstration of  learningface.ai and its use in learning engineering as a resource for learning engineering components.

Jim Goodell1

1 INFERable, a Public Benefit Corporation and IEEE Learning Technology Standards Committee; jim@inferable.app

Abstract. 

Keywords: AI, Learning Engineering, Instrumentation. Modular Open Systems Architecture, Design Patterns, Standards, Machine Learning, Models.

Introduction

We developed LearningFace.ai an alternative to general purpose sites hosting model cards like huggingface. The site is designed as a public good resource to make learning analytics specific AI models more accessible to learning engineering teams. It is intended to offer a bridge between research and practice, to empower researchers to get their work published and give learning engineers access to those discoveries for data-driven decision-making to optimize learning. The beta open access site is piloting a new approach to democratizing AI learner state detection as “learning engineering components”, packaged in a form that can be implemented in various learning systems. This poster shown in figure 1 shows learningface.ai in its current state with placeholder learning cards.

Figure. 1. 

Poster showing learningface.ai for hosting model cards for learning analytics and learner state detection models.

Discussion and Conclusion

More research is needed on the generalizability of models made available as learning engineering components as well as more translation of research to formats that could be used by practitioners via learningface.ai. We invite the research community to work with us to bridge the research-practice gap for learning analytics and learner state detection methods and models by publishing learning cards to the site and using the site as a resource.

References

  1. Goodell, J., & Kolodner, J. (Eds.). (2023). Learning engineering toolkit: Evidence-based practices from the learning sciences, instructional design, and beyond. Routledge.
  2. IEEE. (n.d.). IEEE P2247.2 Draft Standard for Adaptive Instructional Systems Interoperability.
  3. IEEE. (n.d.). IEEE P2997 Draft Standard for Enterprise Learner Record (ELR).
  4. IEEE. (2023). IEEE 1484.20.3-2022: IEEE Standard for Learning Technology: Data Model for Sharable Competency Definitions (SCD).
  5. IEEE. (2024). IEEE 1484.2-2024: IEEE Recommended Practice for Learning and Employment Record (LER) Ecosystems.
  6. IEEE. (2025). IEEE 2881-2025: IEEE Standard for Learning Metadata Terms (LMT).
  7. IEEE. (2023). IEEE 9274.1.1-2023: Standard for Experience API (xAPI).
  8. INFERable.app. (n.d.). INFERable.app [Website].
  9. learningface.ai. (n.d.). learningface.ai [Website].