You can find more information about my professional experience on my linkedin or stackoverflow profiles but here are the highlights.
I have completed a PhD in particle physics in 2014 during which I was part of the exciting ATLAS experiment at CERN. Analyzing billions of data events, we were able to show the existence of a long searched particle, the Higgs boson. Its discovery leads to the 2013 Nobel prize in Physics for the theoreticians who predicted the existence of this particle.
After this PhD, I decided to leave academia to explore diverse fields, still through data analysis and mathematical modeling that helps solve real-life problems. A few years later, I discovered graphs through the world of graph databases and especially Neo4j, and couldn’t stop reasoning in terms of graphs since then.
Since 2021, I have been leading the technical development of SmartGrid, a Canadian startup whose goal is to create a mirrorverse, a digital twin of the real world including peaople and their relationships with the digital assets. I have built the first PoC with a Python backend, a small demo frontend (ReactJS) and deployed everything on GCP.
- Data science & machine learning: python data analysis toolkit ie: numpy, scipy, pandas, scikit-learn
- Create the model (mathematical formulation)
- Implementation (efficient code)
- Validation with simulations (Monte-Carlo techniques)
- Visualization (including maps)
- Graph data analysis:
- Neo4j graph database (certified): create the data model that best suits your use-case
- Graph algorithms: extract information from your graph data
- Machine learning on graphs: apply machine learning predictive methods on graph data (graph embeddings…)
Web development: django, HTML, CSS, JS; webservices (RPC, REST, GraphQL)
Database: SQL (Postgresql, Postgis), graph databases (Neo4j, arangoDB)
Development: python (pytest, sphinx), ReactJS. I am also starting to build some stuffs with Rust.
- Cloud: AWS (S3), Google Cloud (Google App Engine)
- 2019-12: Graphie Award - Community MVP
- 2019-09: Winner of Global GraphHack hackathon 2019 organized by neo4j with the project: neomap. Watch the video demo here.
- 2018-12: Neo4j Certified Professional
- 2018-10: second prize at the Lux4Good hackathon (team made of 4 persons):
- Project about measuring diversity within company’s employees (gender, age, disability…)
- Development of a django website from scratch, visualization with D3.js: github repository
- 2019-05: ‘Exploring Graph Algorithms with Neo4j’ video course at Packt Publishing. Available here.
- 2020-11: I was invited by the WinDSML Meetup group from Bussels to give a talk about:
- Graph Databases for Machine Learning: Youtube
- 2020-10: NODES 2020 Online Conference organized by Neo4j:
- Extending a Knowledge Graph from Wikidata: Youtube
- 2019-10: PyConFR in Bordeaux, France (Content is in French):
From 2021: CTO & co-founder at SmartGrid (remote company)
- 2020-04 / 2021-07: Data scientist & backend developer at Deleev (www.labellevie.com), dark grocery mainly operating in Paris, France:
- Logistics and automation through operation research tools (delivery routes and order dispatch on available drivers based on business constraints)
- Code migration to newer package versions (including Python (3.4 => 3.8) and Django (1.9 => 2.2))
- 2011 / today: Teaching
- 2018- : remote mentor (python, neo4j, data science, sql)
- 2014- : private lessons, maths and physics, preparation for the French A-level equivalent
- 2011-2014: Python practical works for undergraduate students
- 2018-10 / 2020-03 : Lead Data Scientist at Motion-S (www.motion-s.com) (startup in the mobility area, Luxembourg):
- Statistical models for driver road accident risk estimation from driver behaviour, contextual and road statistics data
- Road network model using graph technologies
- Dashboards to present results to prospects (VUE.js)
- 2014-12 / 2018-09 : Data Scientist at effiCity (www.efficity.com/) (real estate company, France) (working remotely):
- Real estate estimation algorithm: 30% improvement compared to existing (estimation accuracy)
- Mathematical formulation with parameter fitting on real data
- Implementation with the python stack (numpy, pandas, scipy, sklearn)
- Validation with simulations
- Push to production (python package, model persistence)
- Regression algorithm (KNN)
- Function minimization (scipy)
- Clustering to identify cities with similar caracteristics
- As always, internal note and code documentation
- Participation to the team projects especially web development (django), including CSS integration and JS (jQuery) scripts
- 2011-2014: PhD in Particle physics (Université Paris Sud/LAL, Orsay, France)
- Search for the Higgs boson and, after its discovery, measurement of some properties (called couplings)
- Hypothesis testing (likelihood) and best parameter estimation with the C++ ROOT framework
- Analyze billions of events on distributed data analysis system (Big data)