Medical Imaging AI
Reconstruction, segmentation, and diagnostic-support models for MRI, CT, and microscopy — built with clinical constraints and regulatory reality in mind.
Feasibility study · Model development · Clinical pilot
Boutique AI consultancy
Medical imaging and computer‑vision AI, from research to production.
One senior data scientist, end to end: the person who designs your architecture also writes the code, trains the models, and ships the system. No handoffs, no knowledge gaps.
- 10+ production ML systems
- Clients from startups to Clalit Health Services
- Research to deployment, end to end
Services
Four ways to engage
Every engagement is hands-on and senior-level, scoped to a concrete deliverable — from a two-week feasibility study to a full research-to-production build.
Computer Vision Systems
Detection, measurement, and inspection pipelines that hold up outside the lab: thermal imaging, multi-camera triangulation, real-time constraints.
Prototype · Production build · Accuracy audit
Applied Deep Learning R&D
Taking a research idea to a working model — architecture selection, training strategy, and honest evaluation before you commit a roadmap to it.
Research sprint · Proof of concept
MLOps & Deployment
Experiment tracking, data management, and deployment pipelines set up so your team can iterate on the work after the engagement ends.
Pipeline setup · Team handover
The scientist who designs your system also builds it
PreDev Consulting is a deliberately small practice: direct access to a senior data scientist with production experience in medical imaging, computer vision, and deep learning.
Recent work includes MRI reconstruction, IVF embryo assessment, CT perfusion analysis for stroke triage, and real‑time industrial inspection — each delivered from first architecture sketch to deployed system.
That continuity is the point of the boutique model: nothing is lost between the person who scopes the problem, the person who writes the code, and the person who answers for the results. They are the same person.
- Computer Vision
- Medical Imaging
- Deep Learning
- Algorithm Design
- MLOps
- Production Deployment
Case studies
Production systems, with the numbers to show for it
Selected work in medical imaging, computer vision, and applied machine learning — each deployed, not just demonstrated.
AI-Powered Embryo Selection for IVF
Deep learning-based automated embryo assessment system integrating segmentation, classification, and temporal reasoning. Multi-modal framework combining U-Net, ResNet, Video Swin Transformers, and Siamese networks for comprehensive embryo evaluation from time-lapse microscopy.
- Medical AI
- Deep Learning
- Video Classification
- +5 more
AI-Powered CT Perfusion Analysis for Stroke
Automated CT perfusion analysis pipeline for acute stroke. Core and penumbra tissue classification in about three minutes — versus 5–10 minutes of expert time — to support thrombectomy decisions.
- Medical Imaging
- Stroke Analysis
- Deep Learning
- +2 more
Real-Time Thermal Obstacle Detection
Real-time vertical obstacle detection using thermal infrared imagery. Hybrid classical and deep learning approach for low-altitude autonomous navigation.
- Computer Vision
- Thermal Imaging
- Deep Learning
- +2 more
AI-Driven Passport Photo Validation
Multi-stage passport photo validation pipeline for ICAO compliance. 98-point landmark detection, geometric registration, and automated cropping at scale.
- Computer Vision
- Facial Landmark Detection
- Deep Learning
- +1 more
Client perspectives
What it’s like to work together
In their words — technical leaders at Agile MRI, Clalit Health Services, and a stealth fashion-tech startup.
“Itamar combines strong technical skills with a practical, results-oriented mindset. He doesn't just build models—he thinks about how to make them sustainable and usable. That kind of foresight and professionalism is rare.”
“Itamar exhibited great communication and creative problem solving, which was particularly impressive given his lack of prior experience in this very specialized domain.”
“Itamar quickly grasped the clinical challenge and focused on the most suitable models. He guided the clinicians through the AI training process, which the team found intriguing and insightful.”
Discuss your project
A 30-minute technical call: your problem, the realistic approaches, and whether this practice is the right fit. No deck, no pressure.
Replies within 24–48 hours on business days.
Book an intro call