Building the stethoscope of the 21st century.
20yo founder-engineer based in Budapest, from Ankara. I build AI diagnostic systems for hospitals and the infrastructure needed to run them in the real world.
Founder of Omniscan Medical. First product: Project Atlas.
Project Atlas -> abdominal CT pathology detection for clinical workflows. 0.983 macro AUROC, 89.5% F1, 99.5% nine-region localization. 1,274 patients, 42,922 annotations, patient-level zero-leakage validation. Paper under peer review.
Omniscan Medical -> the company around Project Atlas. TUSEB 2026-A3 Round 1 accepted. Round 2 submitted on April 5, 2026, awaiting decision.
Hospital Infrastructure -> AI does not deploy into hospitals by magic. PACS, databases, reporting workflows, and procurement all matter. I'm interviewing doctors and IT staff across Hungary, Turkey, and Europe to map the bottlenecks before building the layer on top.
Previously -> NeuroLens with Emir Sürmen. Emir built the Flutter app, server, and ML model end-to-end. I led business, data collection, partnerships, and pitch. We won €10K at a mesh hackerhouse and signed 5 LOIs.
ML/AI: Python, PyTorch, YOLOv11, MONAI, PyTorch Lightning, W&B LLM/Agents: Anthropic Claude, OpenRouter, CrewAI, pgvector Web: Next.js, React, Tailwind, Clerk, Supabase Backend: FastAPI, SQLAlchemy, APScheduler Infra: Docker Compose, Nginx, PM2, Hetzner VPS, Let's Encrypt SSL
- Project Atlas: 0.983 macro AUROC, 89.5% F1, 99.5% nine-region localization
- 1,274 patients, 42,922 annotations, patient-level zero-leakage validation
- TUSEB Round 1 accepted (Program 2026-A3, 400K TL over 18 months if fully approved). Round 2 submitted, awaiting decision.
- Paper under peer review at JIIM (Journal of Imaging Informatics in Medicine)
- 1st place INEC Day (Team Nexus), 1st place mesh hackerhouse (€10K, NeuroLens with Emir Sürmen), 2nd place Agentic Discovery Hackathon (ClaimGraph with David Zagyvai G)
- Stipendium Hungaricum scholar at BME (CS Engineering, 3.5 GPA)
If you're a doctor frustrated with hospital IT, an engineer who wants to work on healthcare infrastructure, or someone who thinks AI should actually work in the real world:
