I'm a Data Science & AI Engineer with a strong foundation from ESIEE Paris (Class of 2024–2025). Driven by curiosity and a love for innovation, I thrive on solving complex problems using data-driven approaches and machine learning techniques.
class LucienLaumont:
def __init__(self):
self.name = "Lucien Laumont"
self.role = "Data Science & AI Engineer"
self.location = "Paris, France"
self.education = "ESIEE Paris - Master's in Data Science & AI"
self.passion = ["Environmental preservation", "AI Innovation", "Sports"]
def get_skills(self):
return {
"languages": ["Python", "R", "SQL", "C"],
"frameworks": ["Next.js", "TensorFlow", "Scikit-learn"],
"databases": ["PostgreSQL", "Vector DBs - Pinecone"],
"ai_tools": ["Mistral AI","Azure AI Studio", "Hugging Face", "LLaMA"],
"deployment": ["Render", "Docker"]
}
def currently_exploring(self):
"""
Current side projects and learning adventures
"""
return [
"🎵 Building a Deezer MCP for music recommendation systems",
"🎓 Creating an Alumni Platform for CPES - Jean-Moulin - Torcy",
"🌱 Exploring how AI can help preserve the environment"
]
def get_learning_status(self):
return "🔄 Always building something new! 🚀"|
🏫 ESIEE Paris |
|
📍 Vancouver, Canada - YourMainGuy Consulting | ⏱️ 6 months
🛠️ Full-stack Development
- Backend services in Python
- Frontend interfaces with Next.js
🗄️ Database & Deployment
- Designed schemas and managed data in PostgreSQL
- Deployed and monitored the entire system on Render
🤖 Advanced AI Integration
- Leveraged Mistral AI's cutting-edge models—Embedding, OCR, and Mistral Large
- Built a scalable R.A.G. (Retrieval-Augmented Generation) pipeline for real-time document querying
📄 Document Processing & Performance
- Engineered regex-based chunking algorithms for optimal document segmentation
- Optimized chunk size and indexing for minimal latency and maximum retrieval accuracy
|
🎯 Event: ESIEE Paris Data Science & AI E4 ML Challenge |
|
🎯 Objective: Fine-tune a GPT-2 model to serve as an interactive cooking assistant
graph LR
A[Python Backend] --> B[GPT-2 Model]
B --> C[Next.js Frontend]
D[LLaMA Models] --> E[Synthetic Q&A Dataset]
E --> B
- Generated synthetic Q&A dataset using open-source LLaMA models on Hugging Face
- Curated and cleaned prompts for realistic cooking scenarios
