Fine-tuning Large Language Models, knowledge distillation, and building efficient real-world medical AI systems that matter.
I'm an ML researcher pushing the boundaries of Natural Language Processing and Generative AI. My work lives at the intersection of cutting-edge research and real-world clinical applications.
Currently focused on fine-tuning Large Language Models, knowledge distillation techniques, and building AI systems that help medical professionals make better decisions. I believe the most impactful AI is the kind that works quietly in the background, augmenting human expertise rather than replacing it.
Beyond the terminal: history enthusiast, hiker, and traveler. The best ideas often come from stepping away from the screen.
Developed a RAG prototype for an XNAT-based PACS system targeting deployment at University Klinikum Erlangen. Processed DICOM and NIfTI CT files to populate PostgreSQL with pgvector, enabling vector similarity search on medical image embeddings.
Dockerized a Vision-Language Model for CT scan report generation at University Klinikum Erlangen. Built preprocessing pipeline for CT-Scans and translation layer for English to German medical reports.
Fine-tuned GRITLM embedding model for evidence-based retrieval generation. Created datasets for CheckThat!2025 challenge in preparation for CLEF 2026.
Implemented 1B-parameter student model from scratch with knowledge distillation from 3B teacher (LLaMA 3.2). Achieved desktop inference at 45 tokens/sec and mobile-ready INT8 quantization under 1GB. Built end-to-end pipeline targeting Snapdragon 8 Elite SoCs.
Reduced data processing time by 50% with Python pipeline processing 10K+ RSS feeds using parallel operations. Achieved 85% accuracy with sentence-transformer embeddings and cosine similarity matching.
Ready to collaborate on cutting-edge AI projects?
Let's build something that matters.