Freelancer profile translated to English.

Description

Full Stack Java/JEE & Angular Design and Development Engineer with 9 years of experience, involved from requirements analysis to production deployment. My technical foundation covers Java, Spring (Boot, Batch, MVC, Data), Hibernate, Struts, JSF, microservices architectures, and REST/SOAP service integration. I have a strong DevOps culture (Git/SVN, Jenkins, CI/CD, JIRA) and cloud expertise (GCP, OpenShift, Azure), with a database specialization in Oracle/PostgreSQL, SQL, and PL/SQL (SQL*Plus).

I have recently specialized in applied AI: RAG (Retrieval-Augmented Generation), LLM, Spring AI, LangChain4j, OpenAI/Azure OpenAI API integration, and conversational agent design (including voice/agentic).

In my latest project, I also handled industrial vision/OCR topics (AI/OCR benchmark, Azure OCR recommendation, barcode reading, Basler camera and webcam integration).

Additionally, I have acquired solid experience in software testing & validation, Oracle Data Integrator (ODI) development, and technical training.

Languages

French
Native or bilingual

Workplace preferences

Can work on-site

Paris (up to 50km), Paris (up to 50km), Toulouse (up to 50km), Lyon (up to 50km), Marseille (up to 50km)

Exertis Azenn
IT & Logistics Distribution / Exertis Azenn
LOGISTICS AND SUPPLY CHAIN
September 2025 - January 2026 (4 months)
Lyon, France
ROLE:
Full Stack Developer (Java/Spring Boot & Angular), AI/LLM reference for
industrial OCR and vision on the project.

CONTEXT:
Migration from Reverso to HUB4R to digitize the handling of
material receptions/returns.
The historical flow, primarily based on scanning (NI/barcode), was enhanced with a
camera + OCR approach to automatically extract information from labels
(NI, SN, item code, MAC, barcode), improve item search, and reduce manual entry.
The functional scope was extended from OBS to OBS + SCF.
On the data side, there was a shift from an Oracle-centric model to a multi-database
architecture: Oracle (CLSAS/Reflex) + PostgreSQL within HUB4R.

Achievements:
Migration from Reverso to HUB4R to digitize the material reception/return flow.
Process modernization: from pistol scanning to an enhanced flow with camera + OCR/AI (extraction of NI, SN, item code, MAC, barcode).
Benchmark and client demos (OpenAI Vision, Azure OpenAI, Azure OCR), with subsequent adoption of Azure OCR recommendation.
Expansion of the business scope from OBS to OBS + SCF (business rules, proprietary mapping, multi-candidate case management), with integration to information systems (Reflex, Daisy, Keycloak).
Integration of a Basler industrial camera (PC webcam POC and industrial camera via Raspberry Pi) + Utilization of the Basler SDK to expose image capture and live streaming APIs, integrated into the application to capture material labels (Java/Spring Boot, Angular, Oracle + PostgreSQL).
Strengthened data architecture: Oracle (CLSAS/Reflex) + PostgreSQL within HUB4R.

Technical Environment:

Java 21, Spring Boot 3 (Web, Data JPA, JDBC, Validation), Spring Security
OAuth2 Resource Server, Keycloak, REST + SOAP/XML (JAXB/Jackson XML),
Angular 18/TypeScript, Oracle, PostgreSQL, ZXing (barcode reading), Azure
OCR (computer vision), Azure OpenAI, Maven, Spotless.
Spring boot Angular LLM Computer Vision Microsoft Azure
LA Innovation IA / Cosmétiques Naturels
AI Solution Architect & Fullstack Tech Lead Engineer
December 2024 - July 2025 (7 months)
CONTEXT:
Development of a revolutionary intelligent voice assistant specialized in natural cosmetics, offering a natural conversational experience with interruptions, similar to voice ChatGPT but enriched with a specialized knowledge base.

Achievements:
- Complete AI Architecture: Design and implementation of a RAG (Retrieval-Augmented Generation) system with a vector database
- Real-time Voice Chat: Integration of OpenAI Realtime API for natural conversation with interruptions and audio streaming
- Spring Boot Microservices: Robust backend with real-time WebSocket and advanced state management
- Modern Angular Interface: Responsive frontend with Web Speech API and real-time audio visualizations
- PostgreSQL Vector Database: Implementation of pgvector for semantic search on the product catalog
- Automatic Correction System: Dynamic update mechanism for prices and product data
- HTML Test Interface: Development of validation and debug tools for voice functionalities
- Audio Optimization: Intelligent management of audio chunks to avoid WebSocket timeouts

Technical Environment:
Java 21, Spring Boot 3.x, Angular 18, TypeScript, PostgreSQL + pgvector, Docker, OpenAI Realtime API, LangChain4j, WebSocket, Web Speech API, Maven, Git, HTML5/CSS3, Audio Streaming

Impact:
- Innovative voice chat solution with contextual AI
- Fluid and natural user experience
- Scalable and maintainable architecture
Java 21 LangChain4j Angular 15
Le Groupe La Poste
Java/JEE & Angular DevOps Design and Development Engineer
HUMAN RESOURCES
May 2023 - Today (3 years and 3 months)
Nantes, France
CONTEXT:

With the rapid evolution of technologies and the growing need to digitize its
services, La Poste has embarked on a large-scale digital transformation. As part
of its digital transformation and to better manage the election process, La Poste
identified the need to modernize its election project. This project is essential
to ensure transparent and efficient management of employee candidacies and votes.
The project aims to migrate the old system to a more modern solution based on
Spring, Angular 15, and Java 17. It also aims to integrate a DevOps approach
to facilitate system updates, deployment, and maintenance.

ACTIVITIES:

- Conduct detailed analysis of the existing system and design the new adapted architecture.
- Write technical guides and installation documents to facilitate the transition.
- Provide technical support for the installation and integration of new adopted technologies.
- Serve as a reference point for resolving technical queries and blocks related to the project or other related areas.
- Develop budget estimates, detailing the cost of tasks and proof of concepts (POCs).
- Organize and lead demonstrations for La Poste representatives, presenting work progress and gathering their feedback.
- Refactor and optimize source code by applying best development practices, prioritizing abstraction and modular design.
- Introduce automated testing with Robot Framework to ensure excellent code quality.
- Implement a CI/CD pipeline using Jenkins, Docker, GitLab, and OpenShift to optimize the development process.
- Ensure the successful deployment of the new system during the annual elections, achieving a notable reduction in errors and an improvement in performance.
Kibana Dynatrace Struts Spring Spring boot Hibernate Spring Data REST Java 17 Java 8 Angular 15 Python Robot Framework Maven Oracle Database Groovy Gitlab OpenShift Jenkins Docker SonarQube Nexus JUnit 5