Julius Mayer

Machine Learning Engineer

PhD Researcher

Hi there 👋

I’m Julius Mayer, a Machine Learning Researcher with a strong foundation in Cognitive Science (BSc) and Intelligent Adaptive Systems (MSc). I currently work as a PhD candidate in the Natural Language Processing (NLP) group at the Institute of Cognitive Science in Osnabrück, supervised by Prof. Elia Bruni, where I research language-driven interaction in physical environments.

Before this, I worked in industry on autonomous driving and face recognition. I have 9 years of work experience in software development for machine learning and artificial intelligence, bridging research innovation with industry application.

I am currently finalizing my PhD dissertation and looking to explore new opportunities in impactful AI. Feel free to reach out regarding open roles or collaborations — I’m always happy to connect.

Highlighted Projects

  • Designed and developed iVISPAR, an interactive multi-modal benchmarking framework, evaluating the spatial reasoning capabilities of vision-language models (VLMs) acting as agents in dynamic, physics-based environments using Unity. iVISPAR supports visual 2D, 3D, and text-based input modalities, enabling comprehensive assessments of VLMs’ planning and reasoning skills.
  • Founded and led MicrocosmAI, a research initiative focused on emergent communication and coordination in embodied multi-agent environments with MuJoCo. End-to-end design, development, and content creation of the MicrocosmAI website. Supervised thesis projects (12 theses), study project management (3 years) and teaching courses.   ArXiv Paper Badge
  • Implemented a computational framework to model cortical spike synchrony, demonstrating that spike synchrony reflects the Gestalt structure of the stimulus which can be interpreted as a mechanism for prior probability estimation.
  • Machine Learning Engineer at Dermalog Identification Systems on face recognition R&D, developing unsupervised face quality prediction, and conducting evaluation and benchmarking.
  • Developed a trajectory planning actor-critic agent for autonomous driving systems and contributed to AI projects utilizing LiDAR sensors as AI Research Intern (Master’s Thesis) at Ibeo Automotive Systems (now MicroVision).
  • Researched human-robot interaction scenarios involving humanoid robots (iCub/NICO) and studied multisensory integration of sound and image data as a Student Research Assistant at the Knowledge Technology Group (WTM), Hamburg. ACM Paper Badge 
  • Developed a distributed learning mechanism by evolving echo state networks via neuroevolution, optimizing a population of embodied agents for foraging and collision avoidance tasks. Demonstrated emergent intelligent exploration, adaptive behavior, and self-organizing intelligence in an agent-based system (Bachelor thesis).  
  • Founded and led MicrocosmAI, a research initiative focused on emergent communication and coordination in embodied multi-agent environments with MuJoCo. Supervised thesis projects (12), study project management (3 years) and teaching courses.   ArXiv Paper Badge
  • Implemented a computational framework to model cortical spike synchrony, demonstrating that spike synchrony reflects the Gestalt structure of the stimulus which can be interpreted as a mechanism for prior probability estimation.
  • Developed a distributed learning mechanism by evolving echo state networks via neuroevolution, optimizing a population of embodied agents for foraging and collision avoidance tasks. Demonstrated emergent intelligent exploration, adaptive behavior, and self-organizing intelligence in an agent-based system (Bachelor thesis).  
  • Designed and developed iVISPAR, an interactive multi-modal benchmarking framework, evaluating the spatial reasoning capabilities of vision-language models (VLMs) acting as agents in dynamic, physics-based environments using Unity. iVISPAR supports visual 2D, 3D, and text-based input modalities, enabling comprehensive assessments of VLMs’ planning and reasoning skills.
  • Machine Learning Engineer at Dermalog Identification Systems on face recognition R&D, developing unsupervised face quality prediction, and conducting evaluation and benchmarking.
  • Developed a trajectory planning actor-critic agent for autonomous driving systems and contributed to AI projects utilizing LiDAR sensors as AI Research Intern (Master’s Thesis) at Ibeo Automotive Systems (now MicroVision).
  • Researched human-robot interaction scenarios involving humanoid robots (iCub/NICO) and studied multisensory integration of sound and image data as a Student Research Assistant at the Knowledge Technology Group (WTM), Hamburg. ACM Paper Badge 
Python C# MATLAB Java C++ Unity MuJoCo TensorFlow Stable Baselines ROS PyTorch RL ML AI JavaScript React Native UI OpenGL Qt Git LaTeX WordPress Notion Obsidian Linux PyCharm

Awards

Research Interests

  • Coordination in multi-agent reinforcement learning
  • Design of physical environments for training and benchmarking of AI agents
  • Language-driven interaction in physical environments

Personal Interests

  • Occasional endurance athlete (Ironman finisher 🏅)
  • Enthusiastic board game and tabletop nerd
  • Asimov’s Foundation-Robot-Empire saga enjoyer
  • One band to loop forever: Architects 🔥