I try to build practical data systems and ML applications with a focus on efficiency, offline capability, and clear business impact. My background spans Data Science, Analytics, and Information Systems.
- Languages: English, French, German (B2/C1)
- Interests: AI for Music and Media, Quantum computing (Qiskit), AI in Medicine/Pharma, Information Systems, Business Intelligence, Robotics
-
PulseFlow AI - Offline Music Recommender & Recognizer
Local, Shazam-style audio fingerprinting and CLAP-based embedding search with FAISS, plus LLM-generated transition notes. Runs fully on device via Ollama/HF small models. -
Real-Time Market Sentiment Dashboard
Dash app that fetches NewsAPI/Reddit, runs DistilBERT and TextBlob sentiment in parallel, and generates short market write-ups with GPT-2. Focus on fast preprocessing and clean UI.
You canalso find these information on my Github page: https://philippe-guerrier.github.io/
For more, see my portfolio (complete information):
https://sites.google.com/view/philippeguerrier/home
- Applied ML to real-world optimization (e.g., menu design and operations) with measurable gains in satisfaction, cost, and revenue.
- Organized AI events and produced post-event documentation and training follow-ups (AI & Law at Dau’IA; SCIS at Sorbonne).
- M.Sc. Data Science - Université Paris Sciences & Lettres - Université Paris Dauphine - PSL
- M.Sc. Data Science in International Strategy & Competitive Intelligence - Panthéon-Sorbonne
- B.Sc. track in Data Science with Economics - Universität Bielefeld
- B.Sc. track in Math/Finance/Economics with Data Science studies - Paris-Saclay
- Offline-first and resource-aware.
- Clear evaluation and removals before adding complexity.
- Small, explainable components; clean interfaces; strong defaults; documented constraints.


