Freelancer profile translated to English.

Description

I am a computer science engineer graduated from ENSICAEN, specializing in artificial intelligence, image, and computer vision. My background has provided me with a solid foundation in programming, algorithms, signal and image processing, machine learning, deep learning, and software development.

During my studies, I completed several projects focused on AI and computer vision. I worked on image classification, object detection, segmentation, and visual data analysis. These projects involved handling image datasets, preparing annotations, training models, evaluating their performance, and analyzing their errors.

I also carried out a project on brain tumor classification from MRI images, comparing different deep learning approaches and using metrics like accuracy, recall, F1-score, and confusion matrices. This work strengthened my skills in medical imaging and model evaluation.

As part of a radiomics project applied to ENT radiotherapy, I worked with medical images and segmentations to extract features, studying the evolution of irradiated tissues and their link to predicting post-treatment toxicities.

My background also led me to develop Python pipelines for automating image processing, structuring datasets, converting annotation formats, and generating usable results. These experiences have given me a rigorous, data-driven approach focused on experimentation and critical analysis, with a particular interest in the practical applications of AI in imaging.

Languages

French
Native or bilingual
English
Conversational
German
Basic

Workplace preferences

Remote only

Primarily works remotely

Pixience
Final Project
February 2025 - August 2025 (6 months)
Toulouse, France
During my internship at Pixience, I worked on several artificial intelligence challenges applied to the analysis of dermatological and dermoscopic images.

My tasks involved the design and evaluation of computer vision solutions for skin lesion detection. I participated in preparing datasets, verifying annotations, analyzing their quality, and setting up pipelines for structured exploitation of images and their annotations.

I also worked on a digital shaving problem, aiming to reduce the visual impact of hair in dermoscopic images to facilitate automatic lesion analysis. This work involved image processing, segmentation, mask generation, and visual evaluation of the results.

Another part of the internship focused on skin lesion segmentation, with the goal of precisely identifying regions of interest in the images. I contributed to experimenting with different approaches, comparing results, and analyzing errors, particularly in complex cases related to image variability, blurry contours, or the presence of artifacts.

I also worked on designing an instance segmentation approach for hair in dermoscopic images. The objective was to isolate individual hair structures to better understand their impact on lesion analysis and improve preprocessing steps.

Throughout the internship, I evaluated the performance of the tested methods using appropriate metrics, qualitative analyses, and visual comparisons. I also wrote technical reports to document experiments, results, observed limitations, and avenues for improvement.
Machine Learning Object Detection Data Science
| Centre François Baclesse
Industrial Project
MEDICAL
September 2024 - January 2025 (4 months)
Caen, France
Project focused on predicting toxicities related to radiotherapy in patients with ENT cancers, in collaboration with the Centre François Baclesse.

The objective was to use medical images, particularly CT scans, to extract radiomic features from irradiated tissues and organs. These descriptors were used to study the evolution of structures before and after treatment, in a Delta Radiomics approach.

The work involved preparing DICOM data, manipulating segmentations, extracting features using specialized tools like PyRadiomics, and then analyzing the obtained variables. The ultimate goal was to assess whether this image-derived information could improve the prediction of radiotherapy side effects compared to models based solely on clinical or dosimetric data.
Université de Kingston
Research Internship
April 2024 - August 2024 (4 months)
Research on GANs, familiarization with DNA sequences and transcription factors, generation of realistic DNA sequences, writing a report for publication.