Automation and AI agents
Useful systems connected to real business tools, with clear inputs, controlled actions and dependable handling of operational data.
Meet the La Refonte team
LoïsPerrigon
Loïs builds and strengthens La Refonte’s digital products and the technical foundations behind client websites. His work brings interface, backend, data and automation into one dependable system, with particular attention to performance, maintainability and what the product needs to do in everyday use.
#PromptMaster
Profile
- Role
- Full-stack AI developer
- Location
- Paris, France
- Languages
- French and English
- At La Refonte
- With La Refonte since 2025
Specialist experience
What Loïs brings to the project
Full-stack development
Front-end interfaces, backend services and API integrations designed as one product rather than a collection of disconnected technical layers.
Web performance
Rendering, page weight and Core Web Vitals improved with attention to both user experience and long-term maintainability.
Systems integration
Websites, forms, data stores and business tools connected without rebuilding mature capabilities that already exist elsewhere.
How Loïs works
Build the simplest system that can stay reliable
Loïs starts with the job the product must perform, then chooses an architecture that remains understandable once the first version is live. Automation is added where it removes repeated effort or improves consistency, not simply because the technology is available.
Understand the operational path
The people, data and tools involved are mapped before a technical flow is designed.
Prefer clear boundaries
Components and services receive defined responsibilities so changes remain local and failures are easier to diagnose.
Measure real behaviour
Performance, errors and important product actions are made observable rather than inferred from a successful deployment.
Automate with safeguards
Automated flows include validation, useful failure states and a clear route for human review when judgement is required.
Key strengths
End-to-end ownership
Can trace a product decision from the interface through the backend and into the data or external service behind it.
Technical simplification
Turns complex requirements into maintainable systems with fewer hidden dependencies and clearer operating rules.
Performance discipline
Treats speed and reliability as product qualities that must be designed, measured and protected over time.
Practical AI judgement
Separates genuinely useful automation from novelty and keeps human control where the task needs context or accountability.
Start a conversation
Speak directly with the people who will do the work.
Share your website, visibility or automation priorities with our senior team.
✦
Free introductory call
30 min · Google Meet
Choose a convenient time. We review your context before the call so the discussion stays concrete.
Finding available times...