Émilie Clain

Do you eat your emotions? Me too.

For a long time, I ate too quickly, without listening to what I was really feeling.

We know that the brain takes about 20 minutes to send the first satiety signals (Source: Harvard T.H. Chan School of Public Health, 2020: “Slow down while eating”).

So I tried timing myself...

But it didn't solve anything: it wasn't just a matter of time.

It was a matter of rhythm, gesture, attention.

I needed a simple guide, during the meal, something to help me slow down and reconnect with my senses.

So I created a small web app: a timer + guiding phrases.

And from the first use, I felt it helped me eat differently.

My name is Émilie, I design innovative tools that simplify everyday life.

If the subject speaks to you, if you recognize yourself, or if you just want to try:

➡️ DigZen is freely accessible:

You can comment, write to me, or just try it.

I love getting feedback.

BudgetScan is a mobile application designed to help consumers manage their budget when shopping in stores. Using embedded deep learning technology, the application automatically detects prices from shelf tag photos, extracts them accurately, and sums them up as you shop, all without an external database, scraping, or account creation. A free, simple, and fast tool, designed to make expense tracking as smooth as a glance at the screen.

And in all stores in France and more...

Exploration of the Nautilus — Interactive map enriched by AI

Complete design and development of an interactive cartographic web application retracing the Nautilus's journey in Twenty Thousand Leagues Under the Seas by Jules Verne.

Technical achievements:

Extraction and structuring of geographical and literary data (32 stops, decimal coordinates)

Text extraction algorithm using textual anchors on a corpus of 900,000+ characters

Collection and structuring of 572 entities (marine species) in JSON

Integration of a local LLM (Mistral 7B via Ollama) into a content generation pipeline, with iterative prompt engineering strategy

Development of a complete HUD interface (HTML/CSS/JS): animations, responsive design, dynamic modals

Interactive mapping with Folium/Leaflet.js, geodesic calculations (Haversine formula)

Continuous deployment via Git/GitHub and Netlify

Technical Stack: Python, JupyterLab, Folium, Pandas, Ollama, HTML/CSS/JavaScript, Git/GitHub, Netlify

Result: Project presented to the Jules Verne Society, demonstrating a hybrid approach between algorithmic extraction faithful to the sources and AI-generated content, with systematic validation and correction of results.

🌐 Demo: the-nautilus-exploration.netlify.app

💻 Source code: github.com/mimiecmoua/nautilus-map