Skills
Technical skills
For each technology: where I have put it into practice (work-study, school, team project or personal project) and, where relevant, the modules taken at Efrei. Courses lay the foundations; work-study and projects put them to use.
AI & prompt engineering
Using AI effectively, and understanding how it works.
- Prompt engineeringWork-studyPersonalSchool
Framing the request (context, examples, steps) and iterating until the expected result, at work as on my personal projects.
- Mainstream AI assistantsWork-studyPersonal
Daily use of most of the assistants on the market, choosing the tool for the task: code, writing, research, analysis.
- AI to code, review and testWork-studyPersonal
Code writing, review and testing sped up by AI, always checking the result.
- Understanding AIWork-studySchool
AI briefings and meetings at SGP, and study of the subject during the BUT: how a model works, its limits and risk of error.
- Efrei
- Generative Artificial Intelligence badge
Front-end
Fast, accessible, polished web interfaces.
- TypeScriptWork-studyPersonalSchool
Pages and components of business applications at SGP, then PlanesView and school projects.
- Efrei
- Web Development
- ReactWork-studySchool
Business applications at SGP; interfaces for Vigipol and the Graphs project.
- Vue 3School
Complete CyberTwin frontend: Pinia, Chart.js, in-house design system.
- Svelte 5Personal
Complete PlanesView interface: HUD, panels, cockpit view.
- Next.jsSchool project · teamPersonal
Wheeltrip web client and former portfolio.
- Tailwind CSSSchoolSchool project · teamPersonal
Interfaces of CyberTwin, Wheeltrip and my former portfolio.
- Three.jsPersonal
PlanesView's 3D globe, terrain and aircraft.
- D3 / Chart.jsSchool
Data visualisations for Vigipol and the CyberTwin dashboard.
- HTML / CSS / JavaScriptWork-studySchool
Fundamentals learned during the BUT then at Efrei, applied to the rebuild of the prefecture's booking site interface.
- Efrei
- Web Development
Back-end & API
Well-structured, tested and documented APIs.
- Symfony / API PlatformWork-study
Evolving services and endpoints of business applications at SGP, with their API tests.
- PHPWork-study
Complete rebuild of the backend of the prefecture's booking site and other internal tools.
- Node.js / ExpressSchool
CyberTwin API, a pair project; stack also seen during the BUT.
- Python (Flask, Pandas)School
Basics learned during the BUT then at Efrei, applied to a route calculator for the Paris metro (Dijkstra, Bellman-Ford).
- Efrei
- Python for Data Science
- FastAPISchool project · team
DataMedia API of the Wheeltrip team project.
- JavaPersonalSchool
Object-oriented programming studied during the BUT then at Efrei, applied to my RPG Ryuukon Palace.
- Efrei
- Java Programming
- Advanced Java Programming
- Full Stack Development with Javain progress
- C# / .NETSchool
Introduction to the language and the .NET platform in the engineering programme.
- Efrei
- .NET and C# Environmentin progress
Mobile
Native Android and cross-platform.
- Flutter / DartSchool project · team
PRM assistance app of the Wheeltrip team project.
- Kotlin / AndroidWork-study
Property inspection app developed for the prefecture (Room, XML).
Data
Modelling, querying and visualising data.
- PostgreSQLWork-study
Databases of SGP's applications: understanding the data model, reading and writing queries for my tickets.
- MySQLWork-studySchoolSchool project · team
Data modelling for the prefecture's booking site, CyberTwin and Wheeltrip.
- Efrei
- Databases
- SQL ServerSchool
Databases of the BUT Informatique.
- MongoDBSchool project · team
Storage of Wheeltrip's journeys and transport data.
- Information systems modellingSchool
Design of data models and information systems.
- Efrei
- Introduction to Information Systems
- Information Systems Modellingin progress
Systems & DevOps
Containers, orchestration, automation and networks.
- DockerWork-studySchool project · team
Daily working environment at SGP; Wheeltrip project infrastructure.
- Kubernetes (k3s)School project · team
Deployment of Wheeltrip on a k3s cluster (ConfigMaps, Secrets).
- AnsibleSchool project · team
Starting and stopping Wheeltrip services with playbooks.
- GitLab / GitHub ActionsWork-studySchool project · team
Merge requests and code reviews on GitLab at SGP; Wheeltrip's GitHub Actions pipelines.
- Linux / WSLWork-studySchool
Development under WSL at SGP, including reporting a memory crash in the tests.
- Efrei
- Operating Systems
- NetworksSchool
Network protocols and how networks work.
- Efrei
- Networks and Protocols
Security & real time
Modules of the second year of the engineering programme.
- Computer SecuritySchool
ING2 year module, in progress.
- Efrei
- Computer Securityin progress
- CryptographySchool
ING2 year module, in progress.
- Efrei
- Cryptographyin progress
- Real-Time SystemsSchool
ING2 year module, in progress.
- Efrei
- Real-Time Systemsin progress
Fundamentals
The basics that make everything else solid.
- C / C++School
Low-level programming, with a small data spreadsheet (CDataframe) in C.
- Efrei
- C and C++ Programming
- Data Structures
- Computer Architecture
- Algorithmics & graphsSchool
BFS, DFS, Dijkstra, Bellman-Ford, Floyd-Warshall, Prim and Kruskal, implemented and visualised.
- Efrei
- Algorithms and Graph Theory
- Optimisation and Complexity
- Kalman filters & geodesyPersonalSchool
Turning 1 Hz positions into smooth 60 fps motion in PlanesView, with tests of WGS84 projections.
- Efrei
- Signal Theory
- ScalaSchool
Functional programming, studied in ING2.
- Efrei
- Functional Programming with Scalain progress
AIArtificial intelligence
AI, a real work tool
It doesn't replace reasoning: it speeds it up, provided you know how to direct it and check what it produces.
Prompt engineering
Frame the context, give examples, break a request into steps, iterate: the quality of the answer depends on the quality of the question.
Several tools, not just one
I can use most of the AIs on the market and I choose depending on the task: code, writing, research, analysis, knowing the strengths and limits of each.
Understanding, not just using
I'm interested in what happens behind the scenes: how a model produces its answer, why it can be wrong, how to check. I review and test what it produces.
An AI culture in the company
At SGP, the AI culture is advanced: regular briefings, meetings, feedback among colleagues. Added to my personal practice and my training, it clearly speeds up my daily work.