Evelity — universal mobile wayfinding app for people with disabilities (Java / Spring)

• Shipped the Evelity back-office end-to-end (React/Vite + Spring Boot 3) — focused on scalability, maintainability and mobile compatibility.
• Designed a multi-datasource architecture with distributed transactions, batch jobs and async messaging.
• Drove quality engineering: code reviews, automated tests, strict CI (coverage ≥ 80%, TDD, SonarQube as a blocking gate), GKE deployment.
• Security: closed OWASP audit findings, hardened Keycloak / OAuth2 / OIDC / JWT auth.
• Steadily paid down technical debt through targeted refactors without slowing delivery.

Evelity Vision — visual SLAM indoor localization system, shipped to production in April 2026 (Python / SLAM)

• Designed and shipped Evelity Vision from scratch in 8 months — full ownership of architecture, back, front and ops.
• End-to-end pipeline: 360° video → frames → Visual SLAM → GPS georeferencing → export.
• Integrated Stella VSLAM (open-source C++ engine) into a Python back-end via a JNI bridge.
• Implemented Sim(2) / Umeyama georeferencing with automatic building-floor detection.
• Async FastAPI back-office + React / Zustand front-end; 2D MapLibre and 3D Three.js visualization.

Consulting missions: job-search management platform (Spring Boot + VueJS) and SMS marketing app (Spring Boot + Angular + MongoDB)

• Built a web platform to help job seekers manage their applications.
• Maintained and evolved an app for processing contact databases used in SMS marketing campaigns.

• Taught programming at bachelor's and master's level.
• Supervised student projects and designed lab exercises.

Applied research in computer vision and deep learning

• Designed and trained deep neural networks for image processing and analysis (B&W / infrared / panchromatic colorization, handwriting-style discrimination, face recognition).
• Built image datasets, evaluated models and tuned hyperparameters.
• 4 IEEE / EUVIP papers published between 2018 and 2020.