Linas Baltrunas

 Netflix, Inc


LinkedIn: link.
E-mail: lbaltrunas (at) netflix (dot) com

Current Research Interests

My main interests are in various (context-aware) personalization techniques. I try to push forward the state of the art in context-aware reasoning, personalized learning to rank, tensor factorization, diversification methods for recommender systems.

I am also interested in variuos aspects of large scale systems that can serve milions of users worldwide.

I love the magic moments of the ideas coming to live.


Our paper on using RNNs for session recommendations was presented at ICLR2016.

I recently joined Neflix, where I continue working on personalization and various machine learning problems.

We released a new data set for context aware RS research: http://baltrunas.info/research-menu/frappe

We released a python recsys serving engine: https://github.com/grafos-ml/frappe/

We presented the tutorial on learning to rank at RecSys'13

Implementation of climf by Mark Levy (mendeley) using graphchi and Zeno Gantner (nokia) using mymedialite.

We released Frappe: a context-aware app recommender for android into the wild.

We got the best paper award at Recsys 2012 pdf !


  • Extensions

    The Joomla! content management system lets you create webpages of various types using extensions. There are 5 basic types of extensions: components, modules, templates, languages, and plugins. Your website includes the extensions you need to create a basic website in English, but thousands of additional extensions of all types are available. The Joomla! Extensions Directory is the largest directory of Joomla! extensions.