Atmane M.

**Mission**: Industrialize the R&D platform and accelerate Time-to-Market.
**Situation**: Research projects were slowed down by the lack of software standards and manual deployment processes, creating friction between design and production.
**Actions**:

• Standardization (Golden Paths): Design Python Starter Kits (Cookiecutter Templates) to automate project structure and ensure a standardized start from day one.
• CI/CD Industrialization: Design GitLab CI pipelines (multi-project, triggers) integrating software compliance: security scans and quality analysis via SonarQube.
• Development & Cloud Native: Design APIs with FastAPI and orchestrate containerized environments using Docker and Kubernetes.

**Results & Impact**:

• Autonomy: Implement a self-service platform adopted by several teams, eliminating dependency on the infra team for each new project.
• Quality: Systematic automation of security and compliance checks on each new project.
• Productivity: Significant reduction in project setup time, representing a saving of several FTEs at the department level.

**Mission**: Develop and ensure the reliability of critical software components for the new Parisian ticketing system (IDFM).
**Situation**: Within Flowbird, on behalf of Comutitres, the challenge was to design an ultra-robust embedded system capable of handling data flows from millions of daily users on the RATP network.
**Actions**:

• Backend Development (Python/Flask): Design high-availability applications for ticketing terminals in a constrained environment.
• IoT & Supervision: Implement MQTT protocols (exchange resilience) and SNMP (hardware fleet monitoring).
• Embedded Linux Expertise: Fine-tuning system under Debian and resolving complex latency bugs to ensure 24/7 service.
• Industrialization: Automate quality control with Jenkins (unit/integration tests) and containerize services with Docker Compose.

**Results & Impact**:

• Reliability: Drastic reduction in software incident rates on automated ticket machines and validators in production.
• Modernization: Successful migration to the new Île-de-France ticketing standard, facilitating fleet maintenance.
• Scale: Deployment of a stable solution supporting massive transaction volumes for a global metropolis.

**Mission**: Design an R&D POC for the reverse engineering and digitalization of dismantling projects.
**Situation**: In the context of dismantling complex installations, technical history is often scattered across decades of unstructured archives. The goal was to intelligently reconstruct this data to make it usable.
**Actions**:

• Data Extraction (OCR): Implement a processing chain to digitize and convert paper archives into text data streams.
• Data Analysis (NLP): Use Spacy and Topic Modeling to identify technical events, key dates, and major interventions in the documents.
• Structuring & Storage: Model a MongoDB database suitable for indexing heterogeneous content.
• Temporal Analysis: Develop algorithms to correlate extracted information and reconstruct a coherent history.

**Results & Impact**:

• Lifecycle Reconstruction: Create an automated historical timeline, allowing tracing the evolution of the installation and past interventions.
• Decision Support: Transform "dead" archives into a chronological research tool for engineers responsible for dismantling.