Raphaël Bournhonesque

Machine Learning Engineer at Open Food Facts
  • Claim this Profile
Contact Information
us****@****om
(386) 825-5501
Location
Paris, FR

Topline Score

Topline score feature will be out soon.

Bio

Generated by
Topline AI

You need to have a working account to view this content.
You need to have a working account to view this content.

Experience

    • France
    • Technology, Information and Internet
    • 1 - 100 Employee
    • Machine Learning Engineer
      • Sep 2022 - Present

      Ville de Paris, Île-de-France, France

    • France
    • Advertising Services
    • 100 - 200 Employee
    • Data Scientist
      • Jul 2019 - Apr 2022

      Paris, France - Developed an image similarity service, using EfficientNet architecture for image embedding and Approximate Nearest Neighbors (ANN) for fast image retrieval. Used on a dataset of 2M images. Triplet loss, EfficientNet, ANN - Designed and developed a hybrid NLG system to suggest responses to social media reviews (Google Maps, Facebook). Using a library of manually curated response sentences, relevant sentences are selected and aggregated in a specific order based on the detection of tags in… Show more - Developed an image similarity service, using EfficientNet architecture for image embedding and Approximate Nearest Neighbors (ANN) for fast image retrieval. Used on a dataset of 2M images. Triplet loss, EfficientNet, ANN - Designed and developed a hybrid NLG system to suggest responses to social media reviews (Google Maps, Facebook). Using a library of manually curated response sentences, relevant sentences are selected and aggregated in a specific order based on the detection of tags in the input review. Tags are detected using fine-tuned RoBERTa/CamemBERT models (en/fr language respectively). NLG, document classification, transformers library Show less

    • France
    • Software Development
    • 1 - 100 Employee
    • Data Scientist
      • Jun 2018 - Jul 2019

      Lyon, France - Improved the relevance of results returned by the internal search engine (based on Elasticsearch). Decreased user search latency by 60% by developing a dedicated NER service. - Trained and deployed a Tensorflow NER model (biLSTM+CRF) to detect person names in search queries. - Designed and implemented an algorithm to detect noun phrases in search queries, using syntactic parsing on the Common Crawl corpus.

    • France
    • Advertising Services
    • 1 - 100 Employee
    • Software Engineer
      • Jan 2017 - Jun 2018

      Lyon, France - Designed and developed a multi-platform (Facebook, Twitter, Slack, web) framework to build chatbots. Python, PostgreSQL, RabbitMQ, ZeroMQ, asyncio. - Developed an intent detection service (document classification task) using biLSTM models with custom French word embedding. Keras, Gensim, ZeroMQ - Trained and deployed intent detection models, deployed a monitoring stack. AWS, Grafana, InfluxDB.

    • Technology, Information and Internet
    • Cofounder
      • Apr 2016 - Nov 2016

      Paris, France ZebrIA is a personal medical assistant that makes the life of chronic patients easier. - In charge of all technical aspects of the product: architecture, development, deployment - Designed a natural language understanding framework in Python to understand user messages and answer appropriately - Promotion 1.2 of French startup accelerator Startup42 (Epita-Epitech)

Education

  • Ecole normale supérieure de Lyon
    M.S. in Biology, Biology
    2012 - 2016
  • Ecole normale supérieure de Lyon
    M.S. in Computer Science, Computer Science
    2014 - 2015

Community

You need to have a working account to view this content. Click here to join now