Biomedical Engineer · MSc Data-driven Health @ KTH · Working at the intersection of AI and healthcare.
Currently annotating medical imaging data at Neko Health and building projects around computer vision, medical data, and health AI.
Gait Analysis Computer vision pipeline for extracting joint and segment angles from raw video — contour detection, cubic spline interpolation, and standardized gait phase breakdowns. No pose estimation library, just OpenCV and math.
WSI Preprocessing Scalable HPC pipeline for preprocessing liver cancer Whole Slide Images — tiling, background removal, and blur detection to filter high-quality tissue regions. Containerized with Apptainer for reproducibility.
Healthcare Nanochat GPT-style medical chatbot trained from scratch on MedQuAD and MedDialog datasets. Custom BPE tokenizer, two-stage training pipeline (pretraining + fine-tuning), safety evaluation suite with 30 curated test cases, and SLURM/Bianca cluster support.
Python OpenCV NumPy pandas scikit-learn PyTorch matplotlib Jupyter