Gaëtan Pruvost

Machine Learning Engineer

Orsay, Île-de-France, France

About

My new quest: automate knowledge extraction for system & software engineers. After a few years in providing model-driven tools to design real-time and embedded systems, I moved back to my first passion: Artificial Intelligence. With a new challenge ahead: push innovation from the idea to the real world.

Experience

  • Thales (Permanent · 13 yrs 7 mos)
    • Data Scientist - Machine Learning Engineer
      Sep 2021 - Present · 4 yrs 10 mos

      Applied Research on deep neural networks with applications to: - Natural language processing for design assistants. - Signal processing for defence and security domains. Implementation of neural net architectures in Pytorch (+Lightning). Contribution to both applied research projects and prototyping of innovative solutions (from theoretical evaluations to real-world applications).

    • Software Architect - Artificial Intelligence / NLP
      Jan 2018 - Sep 2021 · 3 yrs 9 mos

      Definition and realization of AI services to enrich engineering workbenches. Use of NLP (bert embeddings, semantic search, topics extraction,...) and data analytics to implement a smart assistant for system & software engineers. - Prototype implementation, internal startup mode (pitchs internes, hackatons...). - Recruiting and management of a hybrid team of data scientists & developpers. - Design authority for NLP & architectural choices.

    • Technical Manager - Code generation for embedded systems
      Jan 2015 - Dec 2017 · 3 yrs

      Driving the industrialization of research results for embedded systems engineering. Application of MBSE (modeling based engineering) and automatic code generation to real-time and embedded systems. Deployment in operational conditions. - Technical management of the engineering/development team. - Pre-sales interactions with customers: demonstrations, customizations...

  • PhD student & Research studies engineer at LIMSI-CNRS (now LISN)
    Sep 2008 - Dec 2012 · 4 yrs 4 mos

    Modelling of Human-Computer Interaction (HCI) in pervasive computing and development of an automatic reasoning engine to support the continuous adaptation of user interfaces to the context of use. http://www.theses.fr/2013PA112017

  • Assistant teacher (ATER) at IUT d'Orsay
    Sep 2011 - Oct 2012 · 1 yr 2 mos

    Supervision of practical sessions for computer science students at the university - Hardware architectures (building/simulating a processor) - System programming

  • Instructor (Moniteur) at Supelec
    Sep 2008 - Aug 2011 · 3 yrs

    Supervision of practical sessions for Supelec students in computer science - Multithreading - Basic algorithmics

  • Intern at INRIA
    Feb 2008 - Aug 2008 · 7 mos

    I did my master thesis on Gesture Recognition with a dataglove. Gesture recognition was achieved via a new approach for recognition I proposed during the thesis which mixes petri nets and fuzzy logic. It was applied to an interactive application in the field of molecule visualization.