Our Mission

CEPIA leads AI-driven pedagogical research and innovation within the IONIS Education Group, uniting researchers, doctoral students, and engineers across its technology-focused schools. Our goal is to use artificial intelligence to better understand each student, anticipate their challenges, and act proactively to deliver more personalized education for everyone. We design, test, and prototype AI-powered educational solutions, combining scientific rigor, operational agility, and a commitment to ethical standards. As a hub for applied research and innovation, CEPIA develops prototypes, personalizes learning pathways, and ensures transparent AI integration. We bridge General Directors, Research, and Pedagogical Directors to drive coordinated transformation, supported by continuous scientific and technological monitoring.

Our Impact

CEPIA’s ecosystem, built on decades of pedagogical expertise, aims to drive measurable outcomes:

  • Scale: Serving students and educators across IONIS schools, with multiple active research projects.
  • Achievements: Since 2025, solutions are being prototyped and tested internally at EPITA and EPITECH.
  • Goals: By 2027–2028, the ambition is to reach full student coverage, high user satisfaction, reduced dropout rates, and international recognition as a leader in AI-enhanced education.
  • Partnerships: Collaborates with MOCAH-LIP6 (Sorbonne/CNRS) for adaptive feedbacks, TALENT-IRIT (Toulouse) for learning technologies, and Microsoft Azure AI for advanced computing and GDPR-compliant infrastructure. The strategic roadmap includes scaling solutions to all IONIS schools, adapting to public education, and establishing a global standard for AI-driven learning.

Research and Directions

At CEPIA, our mission is to use artificial intelligence to better understand each student, anticipate their challenges, and act proactively to deliver more personalized education for everyone. We unite researchers, doctoral students, and engineers to design, test, and prototype AI-powered educational solutions with scientific rigor, operational agility, and a commitment to ethical standards. This motivation guides all our research directions and underpins our commitment to transforming education through responsible and innovative AI.

Our Mission and Motivation

Education is undergoing a profound transformation driven by advances in artificial intelligence and data science. At CEPIA, we believe that AI and machine learning can help address some of the most pressing challenges in education—such as personalization, accessibility, and the need for lifelong learning—while also raising important questions about ethics, equity, and human development. Our research lab is committed to exploring these opportunities and challenges with rigor, responsibility, and a focus on real-world impact. We aim to bridge the gap between technological innovation and pedagogical practice, ensuring that new tools and methods truly benefit learners and educators.

Our Solutions

Autonomous Assessment

Currently in internal prototyping at EPITA and EPITECH, our AI-powered system automates code project grading and provides real-time, personalized feedback. It identifies common errors and suggests improvements via dashboards for teachers to better understand student challenges. The system supports multiple programming languages, adapts to different grading rubrics, and enables teachers to review and override automated feedback, ensuring transparency and pedagogical alignment. Ongoing work focuses on expanding to written assignments and integrating peer review features.

Potential benefits include reducing the time teachers spend on grading, increasing the consistency and fairness of assessments, and providing students with faster, more actionable feedback. This can help educators focus more on mentoring and less on repetitive tasks, while students receive clearer guidance on how to improve their work.

Predictive Dropout Analysis

Under development, this tool uses machine learning to analyze engagement data and identify at-risk students. It is currently being tested internally to trigger alerts and recommend interventions. The system aggregates data from LMS activity, assignment submissions, and attendance records to build risk profiles. Future work aims to refine predictive models, incorporate additional data sources, and provide actionable recommendations for both students and staff.

By identifying students who may be struggling early, this tool can help educators and support staff intervene proactively, offering guidance or resources before issues escalate. The goal is to foster student retention and success through timely, data-informed actions.

Pedagogical AI Assistant

Our chatbot agent, integrated with course content, is being prototyped to support student queries and offer contextualized answers. It is available 24/7 in test environments and tracks interactions for continuous improvement. The assistant can answer questions about course material, provide study tips, and direct students to relevant resources. Planned developments include integration with external knowledge bases, multilingual support, and advanced analytics for educators to monitor student engagement and common challenges.

The assistant has the potential to increase student autonomy, provide immediate support outside of class hours, and help students overcome obstacles as they arise. For teachers, it can highlight frequently asked questions and common difficulties, informing future teaching strategies.

Adaptive Content Generation

A pilot tool generates tailored quizzes (MCQs, true/false, open-ended) from course materials, integrated into LMS platforms like Moodle. Quizzes adapt to student proficiency levels and are being evaluated in early trials. The tool leverages natural language processing to extract key concepts, generate distractors, and align questions with learning objectives. Future enhancements include support for multimedia content, question banks for various disciplines, and analytics on student performance to refine content adaptivity.

This approach aims to make learning more engaging and personalized, allowing students to practice at their own pace and focus on areas where they need improvement. For educators, it can simplify the creation of high-quality assessments and provide insights into class-wide learning trends.

Research and Innovations at scale

In partnership with MOCAH-LIP6 and TALENT-IRIT, we are aligning large language models (LLMs) for higher education and working on explainability in automated recommendations. We contribute to open-source libraries and plan to release anonymized datasets and educational AI toolkits. Our research also explores ethical AI, user experience design, and the impact of AI tools on learning outcomes, with findings shared through publications, workshops, and collaborative projects.

These research efforts aim to advance the state of educational AI, promote transparency and trust, and ensure that new tools are both effective and ethically sound. Sharing knowledge and resources with the wider community supports innovation and responsible adoption across the sector.

Join our Ecosystem

Events

  • Webinar: “AI & Pedagogy: Innovation in Action” features demos, user testimonials, and workshops.
  • Podcast: “Pedagog.ia Talks” covers trends and expert interviews on pedagogy and AI.

Training

Our certified program, “Integrating AI into Pedagogical Practices,” is being piloted with educators, offering hands-on workshops and resources tailored to each school’s needs.

Collaboration