Merve Dede

• 2026 – Present
  Research Assistant

  Artificial Intelligence and Machine Learning

  Istanbul Beykent University, Istanbul
• 2025 – 2026
  Research Assistant

  Artificial Intelligence Engineering

  Istanbul Beykent University, Istanbul
• 2024 – 2025
  Research Assistant

  Educational Research Program

  Institute Social, Istanbul

• 2024 – Present
  Master of Science

  Data Science and Big Data

  Yıldız Technical University, Istanbul

  In Progress
• 2020 – 2024
  Bachelor of Arts

  Primary Mathematics Teaching

  Marmara University, Istanbul

  Completed

• Educational Data Mining
  The application of machine learning and statistical modeling to data generated within educational systems, with the goal of uncovering patterns that inform instruction, policy, and system design. EDM treats educational datasets as complex, hierarchically structured sources of evidence requiring methodologically appropriate analytic approaches.
• Learning Analytics
  The measurement, collection, and analysis of data about learners in order to understand and optimize learning and the environments in which it occurs. Learning analytics bridges educational research and data science, connecting findings from large-scale assessment programmes to actionable evidence for researchers and practitioners.
• LLMs for Evidence Synthesis
  The use of large language models with schema-constrained output validation to extract structured evidence from bodies of research literature at scale. This approach enables systematic, auditable synthesis of findings across thousands of studies, producing queryable knowledge bases that support evidence-informed research.
• International Large-Scale Assessments
  Cross-national educational surveys — including PISA, TIMSS, PIRLS, TALIS, PIAAC, ICILS, and ICCS — that generate rich comparative evidence about student achievement, educator practices, and system-level factors. Valid secondary analysis of ILSA data requires careful attention to plausible values, replicate weights, and complex sampling designs.
• Machine Learning Education
  The design, implementation, and evaluation of curricula and instructional approaches that develop machine learning competencies in students who are not computer science majors. This includes pre-service mathematics teachers, education faculty, and other disciplinary contexts where ML is becoming methodologically relevant.
• Explainable AI in Education
  The use of model interpretability methods — including SHAP and LIME — to make predictive models applied to educational data transparent and actionable for practitioners who must make decisions on the basis of model outputs. Interpretability is understood here as a functional requirement rather than a technical nicety.

• How do preservice mathematics teachers develop machine learning competencies?

  BAP · 2026
  Yıldız Technical University

  Designing and evaluating instructional sequences for decision tree classification using CODAP. The project examines how participants connect algorithmic thinking to prior mathematical knowledge, contributing design principles for ML education in teacher preparation programmes.

• Adaptive online statistics education for teacher education students

  SUP · 2024
  Boğaziçi University · Grant SUP 20011

  Design and evaluation of an adaptive online introductory statistics course using computer adaptive testing and CODAP-based data exploration. Addresses the persistent gap in statistical training within teacher education programmes.

I welcome collaborations across AI in Education, Educational Data Mining, Learning Analytics, and Large-Scale Assessment Research.