CLEER (CLearinghouse of Engineering Education Resources) is a curated database of high-quality resources in engineering education. It combines journal papers with conference proceedings and ensures whatever you see is of sufficient quality. This means you can spend more time learning about new methods and research in engineering education, and less time scrolling through papers that might not be up to your standards.

CLEER is designed for engineering teachers, early-career researchers in engineering education, and pedagogical advisors who support evidence-informed teaching and curriculum development, but it is open for everyone to use.

CLEER includes both research and practice papers, that are sourced from various journals and engineering education conferences. You can find more about the database here.

CLEER is developed within the EuroTeQ European University Alliance and hosted by EPFL. The curation framework and technical infrastructure are continuously updated by the CLEER research team.

CLEER is supported by Movetia which is funded by the Swiss Confederation, which ensures that the use of CLEER is completely free for the users. Therefore, feel free to explore it as much as you want.

If you would like to use the CLEER curation system for your personal use or as pre-screening for a conference, contact us here.

The resources are curated by a large language model that uses a prompt developed based on curation criteria set by experts in engineering education research. You can read more about the curation model here.

CLEER filters out low-quality papers using expert-developed curation criteria, but it does not re-review or validate the studies. Additionally, as it is based on a large language model’s output, mistakes can happen (although we do our very best to minimize them). Users are encouraged to critically assess each resource’s relevance and rigor for their own teaching or research purposes.

Yes. CLEER continuously expands its database to include reputable sources in engineering education. If your journal or conference publishes relevant research or practice papers, we would be happy to explore inclusion. You can contact us through here.

CLEER prioritizes peer-reviewed journals and established conference series in engineering education or related disciplines. Sources must provide consistent metadata (titles, abstracts, DOIs) and meet minimum quality and accessibility standards. Additionally, as each new inclusion must be curated by CLEER curation system, we would need full text as well, which we will never return the full text to the user. We only use the full text for curation purposes.

CLEER uses semantic search on abstracts, which means that CLEER “understands” the meaning of your query rather than just matching keywords. First, each paper’s abstract is processed once to generate a numerical representation (an embedding) that captures its semantic content — how ideas relate in meaning and context. These embeddings are then stored in a searchable database.

Then, when you enter a query, CLEER converts your query into the same type of embedding and compares it to the stored ones to find conceptually similar papers. Because the abstracts are embedded only once, searches are computationally efficient and fast, even across large datasets. Indeed, CLEER now only needs to embed your query, not the whole database.

This approach allows CLEER to return relevant results even when different terminology is used — for example, linking “active learning in large lectures” with studies on “peer instruction” or “student engagement.” By combining AI-based embeddings with expert curation, CLEER helps educators quickly find high-quality research that supports evidence-informed teaching.

CLEER uses semantic search, which focuses on the meaning of your query rather than exact word matches. This means the system looks for papers that are conceptually related to what you typed, not necessarily those containing the same keywords. Each paper’s abstract is represented by an embedding—a numerical summary of its meaning—created once and stored efficiently. When you search, CLEER compares your query’s embedding to these stored ones to find the closest matches in meaning. So, while your exact words may not appear, the results are chosen because they address the same underlying idea you searched for.

No, all resources that go through the curation, are only used as prompts and are not training the large language model. Additionally, CLEER uses a closed, locally hosted large language model that does not feed back into the original model. The same is true also for all embedded abstracts and metadata, as well as your search queries.

No, your search queries stay on private EPFL servers and are not fed back into the online models. The same goes for papers that have been curated with CLEER curation – all data stays at EPFL in Switzerland.

Yes. CLEER provides access to bibliographic data and abstracts under fair use for educational and research purposes. Full-text papers remain the property of their publishers. CLEER does not host copyrighted material and links to official sources wherever possible.