Xavier Adiyi

Conception et développement d’une application web intelligente permettant la génération de semelles orthopédiques sur mesure à partir de photos de pieds prises avec un smartphone.

Développement de modèles de vision par ordinateur basés sur YOLOv8 pour la détection automatique de points clés du pied (orteils, talon, voûte plantaire) à partir d’images en vue latérale.

Mise en place d’un pipeline IA complet pour l’analyse biométrique : détection d’objets, segmentation, extraction de mesures précises (longueur, hauteur de la voûte, etc.) et traitement d’images par masques binaires.

Conception et déploiement de l’architecture cloud sur AWS et GCP, incluant la conteneurisation avec Docker et l’orchestration avec Kubernetes.

Industrialisation des workflows ML via CI/CD (GitHub Actions) pour l’entraînement, l’évaluation et le déploiement des modèles.

Collaboration directe avec les fondateurs dans un environnement startup orienté R&D, à l’intersection de l’IA, de la santé, de la fabrication 3D et de l’expérience utilisateur.

I contributed to the development of the DibaCare healthcare platform used to manage hospital operations across doctors, patients, specialties, and schedules. I helped design and build both the web application (Next.js) and mobile app (Flutter), ensuring consistent UI/UX and reliable performance across devices. I worked collaboratively with a multidisciplinary team to maintain alignment between frontend workflows and backend functionality, and performed extensive integration testing to prepare the product for a successful MVP launch.

As part of the platform’s intelligent services, I also developed a Healthcare RAG Chatbot that supports medical FAQs and availability queries. This included designing a semantic search pipeline with OpenAI embeddings and a vector database, and implementing intent classification to route queries (FAQs, availability, consultations). I further extended the system with a basic medical triage capability and applied MLOps practices including CI/CD automation, modularization, and testing to ensure production readiness and operational reliability.

Within the Odoo ERP system, I contributed to the development of web applications designed to deliver a smooth and intuitive user experience. I designed and integrated features tailored to user needs while ensuring system consistency and performance.

In parallel, I participated in the development of robust backend applications using Node.js, improving the overall reliability, modularity, and performance of the system. These developments enhanced business process management and increased internal task automation.

I also deployed multiple containerized applications using Docker and Kubernetes on AWS, ensuring scalability, resilience, and easier infrastructure maintenance.

Finally, I implemented VPS (Virtual Private Server) solutions to enhance server performance, security, and availability, providing a stable and production-ready infrastructure.