Agentic AI systems and workflow automation
Designing and building non-medtech agentic AI workflows with attention to tool use, integration, evaluation, and product fit.
Agentic AI - Software engineering - Product deliveryProject experience
Work grounded in high-stakes AI and practical software delivery.
A selection of domains and project types that shaped the BRTLS Logic profile. Some client and implementation details remain confidential.
Medical imaging AI in regulated product contexts
Contributed to healthcare AI projects involving MRI, CT, cone-beam CT, microscopy, and whole-slide imaging, including technical documentation and QMS work for CE and FDA tracks.
Computer vision - ISO 13485 - Deployment lifecycleDetection, segmentation, and longitudinal analysis
Experience across brain metastasis analysis, aneurysm detection, orthognathic planning, sperm cell detection, and cancer characterization from imaging data.
Segmentation - Detection - 2D/3D imagingCNN-based final infarct prediction in acute ischemic stroke
Doctoral research at KU Leuven focused on understanding final infarct prediction from acute CT perfusion using convolutional neural networks.
Deep learning - Medical image computing - Model behaviorDeepVoxNet2 and applied medical imaging research
Developed and used deep learning tooling for medical imaging, alongside projects in tooth, breast, liver, lung, and stroke imaging.
Python - PyTorch/TensorFlow - Research engineeringHow projects are approached
Clarity before complexity.
Strong AI work starts with the problem framing: what decision is supported, what data is trustworthy, what failure modes matter, and what needs to be measured before anyone believes the system.
From there, BRTLS Logic helps move from experiment to implementation, keeping software quality, maintainability, and stakeholder alignment close to the technical work.