CV

Researcher and technical manager who turns multi-source data and real users into systems people can act on — operator dashboards, user evaluations, anomaly detection, and the EU proposals behind them — with a PhD on inferring cognitive state from eye-tracking and physiological signals. Five years across EU-funded R&D; mentor at civic hackathons.

Contact Information

Name Konstantinos Roumpas
Professional Title Researcher & Technical Manager · HCI & Applied ML
Location Athens, Attica

Professional Summary

Researcher and technical manager who turns multi-source data and real users into systems people can act on. Five years across EU-funded R&D: operator dashboards, user evaluations, anomaly detection, and the Horizon Europe and Erasmus+ proposals behind them. In parallel, a PhD on inferring cognitive state — attention, comprehension, stress — from eye-tracking and physiological signals. Trained as a physicist; strongest at reading a project’s technical big picture, spotting what has been missed, and finding the shorter path to a working result.

Experience

  • 2025 - present

    Athens, Greece

    Researcher / Technical Manager
    Frontier Innovations
    Technical direction and research strategy across multiple EU Horizon-funded projects, with primary responsibility for proposal writing and grant management.
    • Write Horizon Europe proposals (work packages, impact, technical sections, Gantt planning) and complete Erasmus+ proposals end to end.
    • Lead technical direction across concurrent EU-funded projects, coordinating a cross-functional team of researchers and engineers.
    • Set research strategy and translate funded project objectives into working AI/ML systems and deliverables.
    • SCUDO (Horizon Europe) — design the operator dashboards for disruption-resilient urban transport, with use cases in Athens, Bavaria, Amsterdam and Venice.
    • BatteReverse (Horizon Europe) — ran the user evaluation and managed the integration of the app’s modules, working out which data feeds which module and how it is visualised.
    • SMARTIN — built an Isolation Forest model that flags anomalous vessel paths on the Danube.
    • Serve as technical point of contact for consortium partners across multiple countries.
  • 2024 - 2025

    Patras, Greece

    Researcher & Teaching Assistant
    University of Patras — Software Quality & HCI Laboratory
    Research and technical delivery across Erasmus+ and Horizon Europe projects, alongside undergraduate teaching.
    • Built machine learning models in Python and PyTorch that infer cognitive state from multimodal signals, including eye-tracking, behavioural data and interaction logs.
    • Developed a physics-informed neural network for gaze prediction that outperformed a conventional baseline, raising vertical-gaze R-squared from 0.85 to 0.91.
    • Designed and ran controlled user studies with physiological manipulation checks, including a randomised crossover protocol.
    • Published four peer-reviewed papers and one open dataset; reviewed five manuscripts for two international journals.
    • Teaching Assistant on the Software Quality course, and supervised an undergraduate research project through to a publishable manuscript.
  • 2024 - 2024

    Rotterdam, Netherlands (Remote)

    Learning Experience Designer & Team Leader
    ATERMON B.V.
    • Designed AI-supported learning modules with data-driven personalisation for EU-funded R&D projects.
    • Led design decisions across digital learning, inclusion, accessibility and human-AI collaboration workstreams.
    • Coordinated delivery with international consortium partners against fixed reporting milestones.
  • 2021 - 2024

    Athens, Greece (Remote)

    Instructional Designer & Developer
    Asserted Knowledge O.E.
    Joined as the sole member of the development team and grew into leading a team of seven.
    • Grew from sole developer to leading a team of seven engineers and designers, defining process and standards as the team scaled.
    • Contributed to a portfolio of more than 50 educational games and tools delivered across EU-funded projects.
    • Designed and evaluated an AI-assisted course authoring tool, validated with 10 educators using the System Usability Scale and semi-structured interviews.
    • Organised and led educator workshops and Erasmus+ Learning, Teaching and Training Activities across partner countries.
    • Ran learning-analytics data analysis in Python and authored technical documentation for project reporting and audit.

Education

  • 2023 - present

    Patras, Greece

    PhD
    University of Patras
    Computer Engineering and Informatics
    • Human-Computer Interaction, Machine Learning, Cognitive Modeling, Adaptive Systems
    • Software Quality and Human-Computer Interaction Laboratory. Supervisor Prof. Michalis Xenos; committee Prof. Nikolaos Tselios and Prof. Ioannis Garofalakis.
    • Research on physics-informed modelling of oculomotor behaviour, treating deviation from a learned reconstruction of gaze as a measure of cognitive state.
    • Applications span attention lapse, reading comprehension, blink-based input, operational stress and VR cybersickness.
    • Eight master’s-level courses completed within the doctoral programme. Expected defence 2027.
  • 2016 - 2022

    Thessaloniki, Greece

    BSc
    Aristotle University of Thessaloniki
    Physics
    • Focus: computational physics, data analysis, non-linear systems.
    • Foundation for the physics-informed machine learning in my doctoral work.

Hackathons & Mentoring

Apps4Athens 3.0 Gov Hackathon · City of Athens, 2026: Mentor for teams building open-data, digital-governance and AI solutions for the city
3rd Social Hackathon & Social Innovation Lab · 2023: With team tinyTerrain (health and safety through gamification) — one of five teams awarded at the Thessaloniki final, after a five-month acceleration programme
Coding Challenge for Ukraine · EDFA, 2022: Participant in a pan-European online hackathon for a more resilient Ukraine and Europe

Projects

  • Gaze Reconstruction as Cognitive-State Measurement

    Physics-informed neural network in PyTorch that reconstructs expected gaze from eye-tracking data. Deviation from the reconstruction serves as a measure of attention lapse rather than as tracking error.

    • Multi-head architecture embedding oculomotor dynamics as constraints in the training objective.
    • Outperformed a conventional neural baseline; vertical-gaze R-squared 0.85 to 0.91.
    • Published in Expert Systems with Applications (2025).
  • CourseCrafter — AI-Assisted Course Authoring

    Authoring tool that lets educators build Minecraft Education courses without coding or prior familiarity with the platform’s controls.

    • Evaluated with 10 educators across geographies, grade levels and subjects.
    • Mixed-methods evaluation using the System Usability Scale and semi-structured interviews.
    • Design published at WAIE 2024; evaluation published at AICCONF 2025.
  • Blink-Based Hands-Free Input

    Method separating voluntary from involuntary blinks so that blinks can serve as a reliable input channel, validated as robust under induced stress.

    • Two-component Gaussian mixture model over blink duration establishes the classification boundary.
    • Stress manipulation validated physiologically via pupil diameter; the boundary held.
    • Open dataset and model published on Zenodo.
  • Research Dashboards & Operational Interfaces

    Interface design work turning dense, multi-source data into something people can read and act on, across EU research projects and operational contexts.

    • SCUDO — dashboard unifying multi-source EU project data into a single readable view.
    • Search-and-rescue flight map, designed for legibility under time pressure.
    • Adaptive lecture platform for a relativistic quantum mechanics course.
    • All designed in Figma.

Publications

Skills

Programming: Python (PyTorch, scikit-learn, NumPy, pandas, SciPy), R, SQL, MATLAB, Wolfram Mathematica, SPSS, LaTeX
Machine Learning: Deep learning, physics-informed neural networks, predictive modelling, explainable AI (SHAP), multimodal data processing, feature engineering
Research Methods: Experimental design, statistical analysis, cross-validation, permutation testing, randomised crossover designs, usability testing, user interviews, System Usability Scale
Sensing & Signals: Tobii Pro Lab, VR eye tracking, EEG (MNE), electrodermal activity, pupillometry, blink detection, fixation and saccade analysis
Design & Tools: Figma, dashboard and interaction design, data visualization, Git, Jupyter, Power BI, Microsoft Azure, Moodle

Reviewing & Evaluation

Horizon Europe · European Commission: Registered expert evaluator
International Journal of Human-Computer Interaction: Peer reviewer — one review completed (2026)
Cognitive Systems Research (Elsevier): Peer reviewer — two reviews completed (2026)
International Review of Research in Open and Distributed Learning: Peer reviewer — one review completed

Certificates

  • Certificate of Reviewing - Taylor & Francis (2026)
  • Certificates of Reviewing - Elsevier (2026)
  • Certificate of Competency in SQL - LearnSQL.com (2025)
  • Prepare Data for Analysis with Power BI - Microsoft (2025)
  • Azure Fundamentals & Azure App Service - Microsoft (2025)
  • Moodle Developer Basics - Moodle (2024)
  • Foundations of User Experience and Design Thinking - Microsoft (2023)
  • Minecraft Certified Mentor & Trainer Academy - Microsoft (2022)
  • Certificate of Proficiency in English (ECPE) - University of Michigan (2012)

Languages

Greek : Native speaker
English : Full professional proficiency — Certificate of Proficiency (ECPE), University of Michigan