Chakir MAHJOUBI (CHAKIR)

En résumé

Entreprises

• Samsung Uk LTD - Computational Linguist

  Informatique | Staines 2017 - 2019 NLU: Checked for natural language responses developed by third party creative content writers. Building data pipelines for the dialogue system. Collects content for training purposes (e.g., text, graphics, animations, audio, glossary, etc.). Discussed data and models with research and development team to make the best possible products for consumers. Applied Named entity tagging, Text sentiment annotation, Voice emotion tagging. Writing tailored language model context-Free Grammars (CFGs), reviewing and maintaining existing ones. Classify and analysed user’s data and investigate intents/goal failures. Investigate ambiguous tagging of the training, potential gaps, proposed recommendations and commit new changes. Maintained the lexicon database. Executed regression test cases for all existing Bixby capsules. Validate and upgrade the software version. Create algorithms to extract information from large data sets Text collection. Developing a proof-of-concept to resolve homophones ambiguity issue, based on the use of WordNet lexical ontology.
  
  ASR: Analyses system output ensuring that phonemes are matched with correct transcriptions, report incorrect output and suggest a solution to resolve the issue. Labelling training datasets for the performance and accuracy of the predictive model. Augmenting domain-specific lexical databases. Attended international conferences and meetings.
  

Formations

• UFR SCIENCES DES LANGAGES DE L’HOMMES ET DE LA SOCIETE

  Besançon (25000) 2000 - 2006 Master’s Degree (MPHIL) in Archives, Document Management and Text Classification.
  

Réseau

Pas de contact professionnel