Pharm.D. · Clinical Research Associate · Medical AI

Clinical Researcher & AI Developer.

Pharm.D., Clinical Research Associate at Hôpital Saint-Antoine (AP-HP), Paris. Independently built a validated AI tool for digestive endoscopic anomaly detection. Bridging pharmaceutical science, clinical research, and machine learning.

GCP ICH E6(R3) Microsoft AI & ML AI Tool — Validated AP-HP Paris, France

View AI Tool Get in Touch

Institution

AP-HP

Hôpital Saint-Antoine · Paris

Availability

Sept. 2026

M2 Alternance

Scroll

01 / Background

Clinical Researcher.
AI Developer.

I hold a Doctor of Pharmacy (Pharm.D.) and work as a Clinical Research Associate (Attaché de Recherche Clinique) at the Endoscopy Department of Hôpital Saint-Antoine (AP-HP), under Professor Xavier Dray, Chef de Service d'Endoscopie Digestive, affiliated with Sorbonne Université.

In parallel, I independently designed and developed an AI-powered anomaly detection system for digestive endoscopic imaging — validated by the senior physicians of the service to assist gastroenterologists in real clinical workflows on PillCam SB3 recordings.

My work combines deep learning (PyTorch, EfficientNet, UNet, MedSAM) with a thorough understanding of clinical research constraints, regulatory requirements, and pharmaceutical quality systems. Fully autonomous, proactive, and experienced in international settings.

Actively seeking an M2 alternance opportunity starting September 2026.

Current Position

Attaché de Recherche Clinique

Hôpital Saint-Antoine · AP-HP, Paris
Since January 2026

Supervisor

Pr Xavier DRAY

Chef du Service d'Endoscopie Digestive
Sorbonne Université

Education

Doctor of Pharmacy (Pharm.D.)

France

Certifications

✓ GCP ICH E6(R3) ✓ Microsoft AI & ML

02 / Endorsement

Letter of
Recommendation.

AP-HP · Sorbonne Université

Hôpital Saint-Antoine · Centre d'Endoscopie Digestive · N° finess: 750100091

Pr Xavier DRAY

Professeur d'Université · Praticien Hospitalier · Chef de Service

Il s'est d'emblée parfaitement intégré à l'équipe. Son assiduité au travail est irréprochable. Il est pleinement impliqué dans les tâches que je lui ai confiées, non seulement dans le projet index, mais plus généralement dans le développement de notre recherche clinique. Je dois même souligner qu'il fait preuve d'initiatives heureuses dans la recherche de nouveaux outils, ou de nouvelles organisations.

Professor Dray highlights Ali's irreproachable work ethic and full commitment to both the AI research project and the broader clinical research activities of the service. He notes his proactive initiative in developing new AI tools that significantly accelerated the research workflow, as well as his strong level of English appreciated in the context of international collaborations.

"Je soutiens pleinement sa candidature" — Professor Dray fully and unconditionally endorses Ali's application and remains available for any further information.

Paris, 17 mars 2026

Verified · Signed Document

03 / Featured Project

AI Tool for
Capsule Endoscopy.

Validated Clinical Tool

Multiclass Anomaly Detection — PillCam SB3

An end-to-end deep learning pipeline for Video Capsule Endoscopy analysis. Performs multiclass segmentation of intestinal pathologies — érosions aphtoïdes, ulcérations, angioectasias — with dual-class contour overlays directly on PillCam footage.

PyTorchEfficientNet-B3UNetMedSAMOpenCVFocal+DiceTTA

Clinically validated by Professors of the Endoscopy Service at Hôpital Saint-Antoine (AP-HP). Directly assists gastroenterologists in PillCam SB3 reading sessions.

Érosions x2 — Dual contour overlay AI Active

Segmentation Pipeline

Angioectasia — Before / After

Extracts frames from long-duration MPG PillCam recordings, classifies pathology type, and generates precise masks. Side-by-side: raw endoscopic image vs AI-annotated output with contour overlay on an angioectasia lesion.

ConvNeXt-SmallSwinV2-SmallMixUpStratified SplitsGoogle Colab

A MedSAM-based annotation tool (HTML canvas widget in Colab) enables clinicians to build tricolor mask datasets efficiently — making the pipeline clinician-friendly from the ground up.