Melinda Xiao-Devins

Melinda Xiao-Devins has over 20 years of work experience. In 2000, they started as a Staff Software Development Engineer at Harmonic. From 2010 to 2017, they worked at Cisco Systems as a Technical Lead and Software Development Engineer. During this time, they designed and implemented IoT, cloud based web server software, utilizing Python, Ruby on Rails, and Java. Melinda also led a team of 6 software developers to design and implement C# .NET WPF based application software for Cisco System's video surveillance system. From 2017 to 2019, they worked at Ellie Mae as a Sr. Engineer, Data Science. During this time, they developed a loan document text classifier using multichannel CNN, LSTM, and pre-trained GloVe embedding, and utilized Keras, Scikit-Learn, and NLTK. Melinda also built a VGG16 based image classifier using transfer learning, to classify mortgage loan document images, and implemented image preprocessing software using OpenCV, and extract visual features for loan document images. Since 2019, they have been working at Zoom Video Communications as a Sr. Manager in AI and NLP. During this time, they hired and managed the first Zoom NLP and AI team, and lead ML engineers across 4 different time zones, to invent and build NLP solutions for Zoom IQ, Zoom meeting, Zoom phone, and other Zoom applications, as well as a SparkML recommendation engine and big data ETL pipeline for OnZoom/Zoom Events. Currently, they are the Chief AI Architect at Visla, leading the AI team and working together with software engineering teams to build AI driven applications for video creation and video editing.

Melinda Xiao-Devins holds a MSEE in Computer Software Engineering from Purdue University, a BS in Physics from Peking University, and is a PHD candidate in Physics from Purdue University. Melinda also holds a Deep Learning Nanodegree from Udacity, obtained in March 2017, and an AWS Data Lab Participant certification from Amazon Web Services (AWS), obtained in July 2021.