Skip to content
EXPERIENCE

EXPERIENCE

Main Profile

  • Deep expertise in AI and Deep Learning, with a focus on Computer Vision and LLMs. Hands-on experience across the full ML lifecycle: data gathering, labelling, augmentation, model training, evaluation, and dataset generation using the HuggingFace framework.
  • Proficient in Python and its core ML stack: PyTorch, Pandas, and NumPy.
  • Builds web projects for fun and practice. Comfortable with HTML, CSS, JavaScript, React, and Astro. Prefers writing plain CSS over reaching for a framework.
  • Computer Science degree with broad programming knowledge, covering most major languages, algorithms, and data structures.

AI Team Manager

Safetrace Visit Website

  • Leading a small research team focused on applied AI, coordinating tasks, running meetings, and keeping the project moving forward
  • Implemented the core model training pipeline from scratch, using XLM-RoBERTa for token classification on our own dataset
  • Contributed directly to building the training dataset, gaining hands-on experience with the full data lifecycle from generation to labelling
  • Gained practical experience in training monitoring, evaluation, and hyperparameter optimization throughout the project
  • Learned how to balance individual technical work with team coordination in a small, fast-moving research environment

Machine Learning Engineer

HUN-REN SZTAKI Institute for Computer Science and Control, Budapest Visit Website

  • Worked on automating coronary artery calcification measurement from cardiac CT scans, with the goal of supporting clinical diagnosis
  • Applied TotalSegmentator for medical image segmentation and explored multiple methodologies in a research-stage environment without a predefined solution
  • Conducted exploratory data analysis on patient data and researched possible approaches to work around the limited dataset size
  • Processed and visualised medical imaging data using Slicer 3D, point clouds, and Python-based tools
  • Gained insight into research methodologies, digital health, and medical document processing while working embedded in a research institute

Research Intern

Max Planck Institute for Biological Intelligence, Munich Visit Website

  • Completed my BSc thesis on site at the institute, working on an automated tracking system for ruffs filmed in a controlled lek environment
  • Built the full training pipeline for a YOLOv8 object detection model, including data labelling and dataset creation from raw video segments
  • Developed the inference pipeline for detecting and tracking individual birds across video frames, using YOLO detections to prompt SAM2
  • The system outputs the location and ID of each bird according to predefined zones, giving biologists structured, usable data
  • Built an interactive Streamlit interface so the biologists could run and explore the tool without any technical background
  • Gained hands-on experience with camera calibration, 2D to 3D point projection, and the full computer vision workflow from raw footage to structured output