Twitch is building the future of interactive entertainment. Recommendations (matching viewers with live content they greatly enjoy), is core to that vision. As an applied scientist intern on Twitch Recommendations Team, you will help design, prototype and implement algorithms and techniques connect Twitch users with the right content. Twitch uses deep learning to perform recommendations. Our approach is currently 30% more effective than using a matrix factorization approach. We are looking for an intern who wants to help us further improve our algorithms. Recommendations is a broad problem, and may include extracting and using data from live video using computer vision (CV), or from chat using natural language processing (NLP). Responsibilities Design, prototype, and implement Machine Learning (ML) recommendation products, leveraging Deep Learning (DL), computer vision (CV) or Natural Language Processing (NLP) Collaborate with team members in the discovery group in an effective manner Requirements Interest in improving content discovery on Twitch via Recommendations Deep knowledge of Deep Learning algorithms and their feasibility for implementation Hands-on experience in developing Deep Learning algorithms, and experience in using Deep Learning libraries, e.g., TensorFlow, MxNet, PyTorch, Theano, etc. Desire and ability to write production quality code Currently pursuing an MS / Ph.D. in computer science or equivalent field Bonus Points Deep knowledge in one of the following areas, with either published research or publicly available code of algorithms Recommender algorithms, e.g., collaborative filtering, content based recommendations, deep models for recommendations, etc. Computer Vision: large scale object detection, activity recognition, OCR, learning with few examples, etc. Natural Language Processing: Word/Sentence Embeddings, Topic Detection, Sentiment Analysis, Entity Extraction, etc. Demonstrated experience in software development via an internship, work experience, coding competitions or submissions to open source projects.