Applied Scientist (fixed Term Contract), Amazon Music AI and Personalization

Amazon Amazon · Big Tech · Tallinn, Estonia · Applied Science

Research scientist role focused on developing novel machine learning solutions for music and podcast recommendations within Amazon Music. The role involves implementing and validating ideas through A/B testing, producing innovative research for peer-reviewed publications, and building scalable models. It requires a PhD or Master's degree and experience in deep learning, ML, or NLP, with a focus on recommender systems.

What you'd actually do

  1. Work backwards from customer problems to research and develop novel machine learning solutions for music and podcast recommendations.
  2. Through A/B testing and online experiments done in tandem with engineering teams, you'll implement and validate your ideas and solutions.
  3. Produce innovative research on recommender systems that shapes the field and meets the high standards of peer-reviewed publications.
  4. Stay current with advancements in the field, adapting latest in literature to build efficient and scalable models

Skills

Required

  • Previous experience building models for business application
  • PhD, or Master's degree in CS, CE, ML or equivalent relevant work experience
  • Experience programming in Java, C++, Python or another modern programming language
  • Experience in neural deep learning methods, machine learning or natural language processing

Nice to have

  • Experience using Unix/Linux
  • Experience in professional software development
  • Experience with modeling and data analysis tools such as Pytorch, scikit-learn, pandas, numpy, scipy etc.

What the JD emphasized

  • novel machine learning solutions
  • innovative research on recommender systems
  • advancing the latest developments in machine learning and AI

Other signals

  • research and develop novel machine learning solutions
  • implement and validate your ideas and solutions
  • Produce innovative research on recommender systems
  • building scalable models