Nlp Research Scientist - One Year Fixed Term Contract M H/F - 75

From a French tech start-up created in 2007, Deezer has become one of the first French unicorns and the second largest independent music streaming platform in the world.

Now listed on, growth is accelerating fueled by #Thepowerofmusic. Deezer is ideally positioned to play a key role in the continued development of the booming music streaming market. If you're looking for an environment where you can grow and have an impact, this is the perfect time to join Deezer!

Our international and passionate teams are spreading the love for music, working to make Deezer the most personal and innovative music streaming service.

We believe music is about diversity & inclusion, multiculturalism and togetherness.

Ready to join the team? Listen: ****Deezer is seeking an NLP research scientist with a strong academic background to join its Research team for a 1-year contract.

The team conducts long-term, exploratory research aimed at advancing knowledge representation of music content and understanding listener interactions with music.

With a catalog of several hundred million tracks and more than ten million users, enabling easy browsing, song retrieval, and music recommendations is a challenge. Your mission will be to develop robust, language-focused, musical item (i.e., songs, artists) representations, with the goal of enabling item or user description generation (e.g., playlist captioning, user preference summary) and supporting user interaction via natural language queries. This will have direct applications in both the recommender system and the search engine of Deezer.

You will collaborate with other researchers, production teams responsible for music recommendation and search, and music curators, and have access to large-scale data on music content and usage, allowing you to have a real-world impact on millions of music listeners.

What you will do:

- Define domain-specific evaluation schemes and collect data;
- Propose, develop, and evaluate NLP systems to create a music item representation space from multimodal data;
- Collaborate with other researchers and with the recommendation and search; team, sharing ideas, knowledge, and projects;
- Scientific watch;
- Scientific communication (papers, publications, blogposts...).