Specialist Dentist · Health Informatician · Clinical Data Scientist — based in Stockholm, Sweden.
I combine 8+ years of clinical care with health informatics and machine learning to build data-driven tools that help clinicians make better, fairer decisions.
- 🧠 Building deep-learning models for medical and dental imaging (MRI classification, dental charting on panoramic X-rays, explainable AI with Grad-CAM)
- ⚖️ Interested in fairness and bias in clinical ML
- 🏥 Working with healthcare interoperability standards — HL7 FHIR, SNOMED CT
- 🌱 Currently exploring clinical decision support systems
| Project | What it does | Stack |
|---|---|---|
| 🦷 Dental Charting on Panoramic X-rays | Numbers every tooth (FDI) and flags caries, deep caries, periapical lesions and impacted teeth on panoramic X-rays; checks Grad-CAM heat-maps for anatomical plausibility and compares the AI with a dentist in a blinded study (in progress) | PyTorch, Faster R-CNN, Grad-CAM |
| 🎓 MSc Thesis — UTI Imaging Decision Framework | Interpretable rules for CT/ultrasound decisions in UTI, mined from 10,758 Karolinska EHR records; flagged 1,300 potentially avoidable scans (~1,800 staff hours) | SQL, Python, Apriori (mlxtend) |
| 🧠 Brain Tumor MRI Classification | Detects brain tumors on MRI — custom CNN vs. transfer-learned ResNet18 & MobileNetV2, 99% accuracy, Grad-CAM explainability | PyTorch, torchvision, scikit-learn |
| 📅 Appointment No-Show & Algorithmic Bias | Predicts missed appointments (110k records) with 11 ML models and audits them for demographic bias using 4 fairness metrics | scikit-learn, imbalanced-learn, Fairlearn, AIF360, SHAP |
| 🩺 UroTrack — UTI Decision Support | Clinical decision support concept for urinary tract infection diagnosis and management (in development) | Python, FHIR, clinical guidelines |
