Machine Learning and Graph Theory

I am interested in modern techniques that combine machine learning and graph theory: Graph Learning, Graph Embedding, Graph Representation, Spectral Graph Convolutions, Convolutional Graph Network, Knowledge Graph, NLP, Computer Vision,… The purpose of these techniques is to exploit the fact that often the data (texts, images,…) are linked. These techniques consist in making a first model of the problem in the form of a graph and then a second modeling in machine learning on the generated graph. Naturally, these techniques find their applications in the fields of:

  • Human resources,
  • Social and organizational networks,
  • Telecommunications,
  • Transport,
  • Pharmaceutical industry,
  • Disease outbreak,
  • Recommender systems,
  • Financial markets,
  • Cybersecurity,

More and more, data are becoming linked, which means that machine learning on graphs is the future of artificial intelligence.

Cloud and Data architecture

I am Togaf enterprise architect and certified in architecting Hybrid Cloud Infrastructure with Anthos. I have 13 years of experience in several industries and 4 years of academic research. In particular, I spent more than 8 years in the hardware department of Bull - Atos, the only manufacturer specializing in the design of supercomputers and data centers in Europe. My knowledge of hardware, firmware and HPC (high performance computing) gives me a great ease with Big Data infrastructure (distributed architectures, on-prem or in the cloud: AWS, GCP, AZURE) and a very good knowledge of new parallelization and pipeline technologies for data and computing such as the Apache Foundation projects (Hadoop, Spark, Kafka, Airflow…). With this experience and drawing inspiration from new Data Fabric concepts, I developed a very integrated vision of Data and Cloud architectures by addressing hardware, network, data engineering, machine learning applications, DevOps and FinOps issues.

FinOps: DevOps and Corporate Finance

FinOps is the new discipline mixing corporate finance and DevOps techniques to control the costs of cloud computing. Indeed, it is a global vision that focuses on the organization of IT teams in order to be able to measure and monitor the expenditure in terms of use of cloud computing by using prediction techniques such as time series or analysis of social networks and of course corporate finance. The goal is to find out when, why, how and by whom a penny is spent on cloud computing. I am interested in FinOps to allow companies to take full advantage of the great potential of cloud computing.

Web Technologies and E-commerce

I am also a fan of web technologies and e-commerce (it happened to me to create an e-commerce website). Due to my interest in the knowledge graph, I am also interested in the semantic web, ontology and taxonomy and graph database technologies in particular Neo4j and Cypher its query language and RDF query protocols such as SparQL. For my blogging activities I often used Html, Css and Flask. And for data science activities, I sometimes do web scrapping to increase and improve data, mainly using Beautifulsoup and Scrapy.

Education
  • International Certificate in Corporate Finance from HEC Paris business school:
    • Financial analysis, Business evaluation,
    • Investment & financing decisions.
  • Specialized Master’s degree in “big data: management and analysis of massive data” from Télécom Paris (Polytechnic Institute of Paris), best engineering school in digital technologies in France:
    • Distributed systems & algorithms,
    • Data engineering & Big Data ecosystem,
    • Big Data security, Cloud computing,
    • SQL & NOSQL Databases,
    • Knowledge graphs & semantic web,
    • Advanced machine learning & statistics,
    • Visualization, Internet of things,
    • Internet economics & personal data law.
  • Ph.D. in graph theory from the University of Paris-Saclay. My thesis subject was to explore algorithmically and topologically the existence of patterns in edge-colored graphs:
  • Master’s degree in operational research, optimization and combinatorics from ENSIMAG, top engineering school in computer science and applied mathematics in France:
    • Graph theory, Algorithmic,
    • Operational Research, Combinatorics,
    • Optimization, Scheduling theory,
    • Probability & statistics,
    • Linear & Dynamic programming,
    • Coding theory, Game theory.

During my career, I have developed the ability to approach the technical and theoretical aspects of my work with a high degree of autonomy and have become comfortable collaborating with multi-disciplinary teams. I enjoy working in an international environment and staying up-to-date with developments in science and technology.