About

Matthew Lightman is an accomplished Senior Machine Learning Researcher with over 12 years of relevant experience in the field. Currently, he is working at Deepgram, where he is responsible for researching new methods for building deep neural network models for audio transcription to text. His goal is to improve accuracy and efficiency in conversational settings where voices may not always be clear, and many people are talking. Prior to his current role, Matthew worked as a Senior Computer Vision and Machine Learning Scientist at Voxel51, where he was the tech lead for the vision team. He was responsible for creating a Learning Framework to train and evaluate computer vision models at scale, using tools such as TensorFlow and Google Cloud. He developed object detection and classification deep models that give insights into video of roads and cities. Before that, he worked as a Data Scientist Lead at JP Morgan, where he and his team were responsible for personalization of the Chase web site and mobile app. They implemented large scale, real-time machine learning systems, which delivered personalized nudges and content that helped customers understand and manage their finances, and gain awareness of product offerings that met their individual needs. Matthew holds a Doctor of Philosophy (PhD) in Physics from Columbia University and a Bachelor of Science (B.Sc.) in Physics from York University. His technical skills include ML, Research Scientist, Computer Vision, Object detection, Neural Networks, Object Detection, TensorFlow, Senior Software Engineer, Data Scientist, Mobile, and Web. Overall, Matthew is a highly skilled and experienced professional who has a proven track record of success in his field. He is passionate about using his expertise to develop innovative solutions that solve complex problems and improve people's lives.