UC Berkeley Cloud Meetup 033: Yoga and Machine Learning (04/28/2022)

Meetup Playlist

Cloud Meetup 033 (Apr. 28, 2022)

Yoga and Machine Learning

“Stay Fit, Stay Safe” is a deep learning solution employing AI to analyze and structure video content. It enables easier discovery and understanding of online yoga classes to help people efficiently practice in the comfort and safety of their homes, while reducing cost. The solution is hosted on Azure cloud platform and includes a user-friendly UI, a highly accurate ML classifier that has been trained to detect over 70 main poses, as well as a recommendation engine, allowing yoga practitioners to specify their preferences and get personalized advice. “Stay Fit, Stay Safe” augments videos with automatically generated summaries, transcripts of the exercises in Sanskrit and English, interpretations of benefits for body and mind, and main contraindications. The program's work has been highly endorsed by certified yoga instructors and “The Art of Living Foundation” for empowering people worldwide to improve their mind and body wellness during the pandemics and beyond.

Dr. Alexandra Savelieva

Dr. Alexandra Savelieva is an applied research lead in Azure Data (Synapse ML) and an alumni of UC Berkeley MIDS program. As part of her hybrid data scientist/engineer role in Microsoft, Alexandra has worked on a variety of interesting problems in Networking and IoT, such as optimization of configurations, analysis of hardware performance, capacity planning, and process mining. She is also driving various projects in "AI for good" area, including multi-disciplinary research in computer vision and NLP for improving the accessibility of online content, enabling privacy-preserving ML inferences, using GANs for enhancing photos, and other innovations that can help people benefit from recent advances of digital technologies in a safe and responsible way. Alexandra is an incoming instructor for the iSchool's Data Science W251 Deep Learning in the Cloud and at the Edge course as of this summer term.

Linda Yang

Linda Yang is a software engineer in Nordstrom Supply Chain Technology. Recently she gained her master of information and data science master's degree from UC Berkeley. Linda leads her team to develop an inbound transportation optimization platform for Nordstrom, enabling better inventory data quality and quantity. In her spare time, she enjoys learning different machine learning models and how to build data pipelines with various cloud infrastructures.