Acerca de Angelo
PhD in Biophysics & AI (Sorbonne/ESPCI/Collège de France), I co-founded Asio Data to bring the high standards of academic research to the corporate world.
❌ My Approach: I do not simply apply off-the-shelf "Black Box" models to your data.
✅ My Method: I seek out the causal and physical rules governing your activity to build robust, explainable, and reliable models.
💡 HOW I CAN HELP YOU:
- 1️⃣ Data Project Audit & Rescue: Your model isn't performing? Your data seems unusable? I intervene to diagnose failures, clean pipelines, and restructure your Data strategy to save drifting projects.
- 2️⃣ Custom Algorithms (Beyond Standard ML): Development of complex predictive models (Bayesian Inference, Physics-Informed Deep Learning) specifically designed for Industry, Biotech, or Finance constraints.
- 3️⃣ Process Optimization & Decision Making: Making noisy data "speak" to extract optimal decision-making levers and reduce uncertainty in your operations.
🛠 Tech Stack:
Python, PyTorch, Scikit-learn, Pandas, NumPy, Git, Docker, Cloud (AWS/GCP).
🎓 The Asio Data Advantage:
Working with me means accessing the methodology of a consulting firm founded by physicists. Rigor, transparency, and documentation are our non-negotiable standards.
📅 Availability:
Part-time or Full-time (Remote preferred / On-site in Paris possible).
👉 Contact me for an initial diagnostic of your specific issues.
Francés
Bilingüe o nativo
Inglés
Competencia profesional completa
Solo teletrabajo
Lleva a cabo sus proyectos principalmente en remoto
Experiencia
- Asio Data (Deep Tech Consultancy)Co-founder & Lead Data Scientist | Scientific Consultingenero de 2026 - Hoy (5 meses)Paris, FranciaCo-founder & Lead Data Scientist | Scientific ConsultingAsio Data (Deep Tech Consultancy)
Mission:
Co-founder of Asio Data, a consulting firm bridging Academic Research rigor and Business Data challenges. We replace "Black Box" approaches with explainable, physics-informed models.My Role (Expertise):
As a PhD in Biophysics & Statistical Learning, I translate complex business ambiguities into rigorous mathematical problems.- 🧠 Advanced Modeling: Constraining AI models with fundamental laws (thermodynamics, symmetries) to ensure robustness.
- 🔍 Bayesian Inference: Extracting weak signals from noisy environments where standard approaches fail.
- 🛡️ Risk Management: Systematically quantifying uncertainty in predictions to secure decision-making.
Services Delivered:
🚀 Operational & Structural Audit:- Diagnosing data quality issues and "cleaning" bias.
- Refactoring fragile manual processes into robust automated pipelines.
💡 Innovation & Custom Algorithms:- Development of bespoke predictive models (Deep Learning, Simulation In Silico / Digital Twins).
- Optimization of complex processes using noisy data.
🎓 Training & Mentoring:- Upskilling technical teams on advanced analysis tools.
- "Data Literacy" training for business stakeholders to understand Data Science potential and limits.
Value Proposition:
"We don't just find correlations; we seek the causes and rules governing your data." - Collège de France - Sorbonne Université - ESPCIPhD Researcher – AI & Biophysics (Physics-Informed AI)Sorbonne Université / ESPCI / Collège de FranceBIOTECNOLOGÍAenero de 2021 - enero de 2025 (4 años)Paris, Francia
Context:
Research conducted within top-tier laboratories (Sorbonne, ESPCI, Collège de France) at the crossroads of Theoretical Physics, Artificial Intelligence and Biology.Objective:
Developing "Physics-Informed" AI models for protein engineering, outperforming classical "Black Box" approaches.Technical Challenges & Achievements:
🚀 Hybrid Deep Learning Architecture:- Design of Neural Networks (PyTorch) integrating strict thermodynamic constraints.
- Result: Ability to simultaneously predict the affinity and stability of protein variants with guaranteed physical consistency.
📊 Big Data Processing & Noise Management:- Analysis of massive sequencing datasets (~100 million reads / Deep Mutational Scanning).
- Applied Bayesian Inference to model and clean experimental noise (over-dispersion, PCR bias) to extract relevant weak signals.
⚙️ Optimization under Constraints:- Implementation of regularization and "Coarse-graining" strategies to enable learning on limited or heterogeneous data, effectively preventing overfitting.
🛠 Technical Stack:Python, PyTorch (Deep Learning), NumPy/Pandas (Data Analysis), HPC (High-Performance Computing), Git. - Sorbonne UniversitéScientific Mentor & Lecturer (Python/Physics) - Sorbonne UniversitéEDUCACIÓN & E-LEARNINGenero de 2021 - enero de 2023 (2 años)Paris, Francia
Description:
Teaching and technical mentoring mission for undergraduate Physics students.- 🗣 Pedagogy & Simplification: Transmitting complex concepts in Thermodynamics and Statistical Physics. Proven ability to adapt technical discourse to various audiences.
- 💻 Digital Project Supervision: Supervision of simulation and programming projects (Python). Code reviews and methodological support.
- 🎓 Group Management: Leading Tutorials (TD) and Practical Labs (TP).
Recomendaciones
Sé el primero en recomendar a Angelo
Ayuda a este freelance a destacar compartiendo tu experiencia.
Estos perfiles de freelance también coinciden con tus criterios
Agatha Frydrych
Backend Java Software Engineer
4.7
(3)
2
Baptiste Duhen
Fullstack developer
4.6
(4)
5
Amed Hamou
Senior Lead Developer
4
(2)
7
Audrey Champion
Web developer
4.3
(3)
4
Formación
- PhD in Biophysics & Machine LearningSorbonne Université / ESPCI Paris / Collège de France2024"Biophysical Modeling of High Throughput Proteins Selection" Research conducted at the crossroads of Statistical Physics and AI within elite laboratories (Collège de France, ESPCI). • Focus: Development of mathematical models and machine learning algorithms to understand and predict protein selection mechanisms. • Key Competencies: Complex Systems Modeling, Statistical Physics, Advanced Deep Learning. (See "Experience" section for technical details and stack).
- MSc in Theoretical Physics of Complex Systems (i-PCS)Sorbonne Université & Politecnico di Torino2021"Biophysical Modeling of High Throughput Proteins Selection" Research conducted at the crossroads of Statistical Physics and AI within elite laboratories (Collège de France, ESPCI). • Focus: Development of mathematical models and machine learning algorithms to understand and predict protein selection mechanisms. • Key Competencies: Complex Systems Modeling, Statistical Physics, Advanced Deep Learning. (See "Experience" section for technical details and stack).