Team lead rapid prototyping projects using Gen(AI) tools, delivering end-to-end projects.

Train image diffusion pipeline that generated novel, unseen 3D shapes 90% faster than conventional manual design methods, leveraging generative modeling algorithms of neural implicit fields.

Build diffusion prototypes using fine tuning, in-painting, knowledge distillation - HuggingFace, Diffusers, AWS, Sagemaker, Gradio, etc..

Apply computer vision models (YOLO, SAM, OWL, Pytorch, MONAI, Vision Foundation Models) for image classification and object detection in manufacturing, life sciences and energy sectors.

Led the integration of AI/ML technologies in cross-functional European teams from 3 different countries formed of engineers, radiologists and surgeons to implement data-driven decisions minimally invasive ablation treatments.

Build statistical, mathematical and ML models using Python, R, and SQL to curate and analyze large, multimodal data including images (CT, MRI) and patient records from European patient clinical trials.

Programmed reproducible and repeatable analytics pipelines for automatic data quality controls of pharmaceutical drugs manufacturing processes.

Developed and implemented efficient data processing pipelines in R and Visual Basic to streamline data analysis procedures, resulting in reduction of validation errors.

Designed and implemented one custom front-end web application for image annotation and labeling (Python, Javascript, HTML and CSS) that sped up the data acquisition and pre-processing.

Applied image processing and statistical methods for quantifying the effect of diet on people's facial appearance with statistical metrics.

Thesis: “Quantitative assessment of ablation treatments for liver tumours — image-based efficacy analysis and predictive modelling”. Parts of my thesis have been published in peer-reviewed journals that you can read about on my Publications.

Thesis: “Image Segmentation and Semantic Description: Tools and Analytics”. Implemented computer vision algorithms for segmentation and classification of skin structures as part of a research project at Philips Research that achieved 98% accuracy with Random Forests.

Thesis: “Volumetric Capnography Respiratory Signals for Spontaneously Breathing Subjects”. Applied signal processing and ML algorithms in MATLAB that achieved classification of respiratory diseases with 90% accuracy.