Do you want your AI research work to impact real data-driven projects ?
And help 80 million active users to communicate more efficiently and precisely in all languages ?

• Reverso is the #1 independent language portal in the world (source: SimilarWeb) with over 80 million active users, 6 billion pages viewed per year.
• The purpose of the organization: Helping the 80 million users communicate more efficiently and precisely in all languages. At Reverso, we apply our AI expertise for their benefit.
• Reverso has more than 20 innovative products: Translation, Translation in Context, Dictionaries, Conjugation, Grammar Checkers, Language Learning Tools, etc. available for Desktop, Web and Mobile, for end users (B2C) and corporations (B2B).
• Our AI-based Grammar Checker corrects 100+ types of mistakes in a click. It offers mistake explanation and a powerful Rephrase feature to take your writing to the next level.
• Our core business is the development of world-class, innovative, and user-centric language products using Deep Learning & NLP. Its 20 + products have one thing in common: they offer state-of-the-art AI technologies, and an innovative UX that seamlessly integrate into the user’s workflow

Job area & mission :

We are searching for a Machine Learning Engineer to optimize our machine learning systems.
You will be evaluating existing machine learning (ML) processes, performing statistical analysis to resolve data set problems, and enhancing the accuracy of our AI software’s predictive automation capabilities.
Within the Tech team, your role will be to work on various projects to improve Reverso tools including translation (CTranslate2), neural-based grammatical correction (NMT or Transformers) and paraphrasing. The objectives are rich and diverse, from optimizing the actual models to the research of novel SOTA approaches up to the testing and launching of the solutions.

Some missions include:

• Creation of complete datasets to enhance the fine-tuning of the models
• Implementation of prefix back-translation to boost the quality and diversity of the generated outputs
• Investigate ChatGPT, GPT3 and other LLMs to transpose the ideas to the Reverso tools.
• Research novel penalty scoring and sampling in beam search.
• Tuning, performance measurement and architecture improvement

Main responsibilities :

• Consulting with Data and Product team to determine and refine machine learning objectives.
• Designing machine learning systems and self-running artificial intelligence (AI) software to automate predictive models.
• Transforming data science prototypes and applying appropriate ML algorithms and tools.
• Ensuring that algorithms generate accurate user recommendations.
• Turning unstructured data into useful information by auto-tagging images and text-to-speech conversions.
• Solving complex problems with multi-layered data sets, as well as optimizing existing machine learning libraries and frameworks.
• Developing ML algorithms to analyze huge volumes of historical data to make predictions.
• Running tests, performing statistical analysis, and interpreting test results.
• Documenting machine learning processes.
• Keeping abreast of developments in machine learning.

Why Joining Reverso :

• Offices in Paris 16/ Neuilly with a friendly co-working environment
• Work on a product that more than 100 million people already use
• Flexible working environment
• Company with a purpose : help millions of users understand, speak and write better in many languages
• International environment
• Attractive salary package

Competencies :

• Master’s degree in Physics, Applied Mathematics or Computer Science.
• Minimum 5 years experience gained in a similar environment with successfull AI applied Research
• In-depth knowledge of mathematics, statistics, and algorithms
• English : fluent
• Proficiency in Python (Java appreciated, C++ and C# is a plus)
• Strong skills in natural language processing, machine learning and optimization are appreciated
• Experience in Pytorch and Tensorflow
• Familiarity with Linux/Unix & Experience with Docker is a plus
• Knowledge in OpenNMT, CTranslate2, GPT, Fairseq & Marian is a plus
• Ability to work independently and strong problem-solving capability