CV

Education, research experience, publications, and skills.

Contact Information

Name Marcel Benjamin Gröpl
Professional Title Machine Learning Researcher
Email merci00@protonmail.com

Professional Summary

MSc in Information Theory and Machine Learning from ETH Zürich, with research experience at Bosch and Sony. Publications at ECCV 2026 (first author) and NeurIPS 2026.

Experience

  • 2026 - present

    Switzerland

    Co-Founder
    Lexy AI
    • Built an LLM-based system for automated legal document generation
    • Ran product demos and pilot tests with law firms, incorporating feedback into development
  • 2026 - present

    Switzerland

    LLM-Bias Study (Co-Lead)
    ETH Juniors
    • Co-led a research study investigating bias in large language models for an undisclosed client
  • 2025 - 2025

    Tokyo, Japan

    Research Intern
    Sony
    • Developed novel NAS algorithms leveraging LLMs as search agents to explore architecture spaces
    • Resulted in a co-authored NeurIPS 2026 paper
  • 2024 - 2024

    Yokohama, Japan

    Research Intern
    Bosch, Computer Vision Division
    • Applied SAM (Segment Anything) to improve pseudo-labeling quality for panoptic segmentation

Education

  • 2023 - 2026

    Zürich, Switzerland

    MSc
    ETH Zürich
    Information Theory and Machine Learning
    • Thesis: Entropy-Gradient Grounding: Training-Free Evidence Retrieval in Vision-Language Models
  • 2019 - 2023

    Zürich, Switzerland

    BSc
    ETH Zürich
    Electrical Engineering and Information Theory

Skills

Machine Learning & Research: Neural Architecture Search (LLM-based search agents), multimodal 2D/3D models (MLLMs), visual grounding, token pruning, LLM fine-tuning (GRPO, SFT), signal processing
Programming & Tools: Python (PyTorch, torchvision, scikit-learn), C++ (graphical computing, rendering)
Computer Graphics: Ray tracing, UV mapping, point-cloud reconstruction

Languages

German : Native
English : Professional
Japanese : JLPT N2
Polish : Conversational
Korean : Basic