What we are looking for:

• A burning desire to solve real world problems at scale by applying Machine Learning
• PhD or Masters in Computer Science, Statistics, or any of the related fields
• 5+ years of research experience with a track record of delivering quality results
• Expertise in machine learning spanning supervised and unsupervised learning methods
• Experience in successfully applying machine learning to real-world problems
• Exceptional problem solving skills
• Strong mathematical skills with knowledge of statistical methods
• Excellent software engineering skills in languages such as Scala, Java, Python, C++ or C#
• Great interpersonal skills
• Strong written and verbal communication skills 

Preferred, but not required:

• Experience as a technical leader
• Experience working with cross functional teams
• Experience in Search, Recommendations, Natural Language Processing, Knowledge Graphs, Conversational Agents, and Personalization, etc
• Experience using Deep Learning, Bandits, Probabilistic Graphical Models, or Reinforcement Learning in real world applications
• Experience with Spark, TensorFlow, Keras, and PyTorch
• Experience with cloud computing platforms and large web-scale distributed systems
• Experience in applied research in industrial settings
• Open source contributions
• Research publications at peer reviewed journals and conferences on relevant topics

Links:

• Netflix Research site
• Our culture
• Our long term business view
Areas:

Page Personalization:

• Our team is responsible for research and development of the machine learning algorithms that create the Netflix homepage for all our over members around the world across all the devices. This is primarily focused on deciding which rows of recommendations produced by our many ranking algorithms to show, so that members can easily find something great to watch and enjoy, but also includes work on in general what all our algorithms should optimize for and how to personalize additional aspects of the user experience. There is very little processing done on the output of our algorithms, so the work we do has a large impact on how people use Netflix. This means that we take on the responsibility for understanding and predicting all of the different reasons that someone may want to visit Netflix and help them find what they need. When we do our jobs well, we help people all over the world find a bit more joy.

Experience Personalization:

• Our team works at the intersection of UX design and machine learning. We research and develop algorithms that decide which information about a movie to show to our members.  For example, we are responsible for selecting the best artwork and synopsis for each show on a Netflix homepage. We also explain why a member should care about a new movie, TV show, or game. The goal is to help our members make great decisions on the next title to watch. A lot of our work involves contextual bandit algorithms including deploying models, investigating and improving state-of-the art-algorithms, and evaluation metrics. When we do our jobs well, we help people to select the best entertainment on the planet.

Personalization Rankers:

• Our core team works on the core recommendations and personalization algorithms. These algorithms (alongside Page Personalization) are responsible for most of what’s shown on the Netflix homepage (which drives 80% of what members in Netflix play). 


Some areas you’ll be working in:
• Recommender Systems and Personalization. Almost every aspect of the Netflix experience is personalized, and much of that personalization is driven by our various flavors of recommendation algorithms.


• Causal Inference And Reinforcement Learning. As we’re ever seeking to align our recommendations with what titles members love, it’s important to tease apart the cause-and-effect of what we’re doing. You’ll work with large scale CI and RL algorithms to achieve this. 


• Large Scale Machine Learning. Netflix is available in over 190 countries, with over 200+ million members. This gives us a unique dataset to work with, but also unique challenges in how we scale our models. You’ll work on cutting edge techniques to scale your models for use in our production systems. 

Search and Recommendations:

• This applied ML team is responsible for improving our member experience by innovating on algorithms for all aspects of Search and other discovery canvases where members express their entertainment needs explicitly. Such mechanisms are a primary way for our members to discover and engage with content on our service. Algorithms innovation in this area, hence is critical, as our content offerings scale and become diverse.
• 
  
• Our work is a mixture of applied research and engineering to do end-to-end machine learning i.e. from inception of an idea to its productization via online and offline experiments.  We work in a highly cross-functional environment, collaborating very closely with PMs, back end and front end engineers, data scientists and engineers, UX Designers, editors, etc. 
• 
  
• Typically we require experience and expertise in Applied ML, Search, Recommendations, Personalization, NLP, KG, HCI, as well as  proficiency in software engineering, familiarity with distributed computing and Deep Learning frameworks. 


Following are some of our recent publications:
• Augmenting Netflix Search with In-Session Adapted Recommendations - RecSys 2022 
• Query Facet Mapping and its Applications in Streaming Services: The Netflix Case Study - SIGIR 2022 
• Recommendations and Results Organization in Netflix Search, RecSys 2021
• Improving Search Results Ranking Using a Knowledge Graph at KINN-CIMK 2021 
• Challenges in Search on Streaming Services: Netflix Case Study SIGIR 2019 

Growth and Notifications:

• Our team focuses on personalization of the end-to-end user lifecycle at Netflix. 
• For example, we work on a better personalized experience for signup (e.g. personalize the signup flow), optimize the payment experience (increase the success of a transaction going through, detect fraud), and personalize the messaging experience (when and what to send via emails, push notifications and pop-ups).


• As Netflix shifts to commerce and ads, we also started shifting our focus to monetization such as detecting sharing and making better predictions of life-time value.


• Functionally, we go from prototyping all the way to launching experiments and productionization. We often  leverage causal inference and context bandits in a lot of our problem formulations. We care deeply about both research and scientific rigor, as well as product impact. We are looking for folk who can act both as a scientist and an engineer.