Liang Li Email and Phone Number

◦ Implement deep learning infrastructures to build custom models for DeepComposer, including creating and maintaining data pipeline, building and implementing networks (GANs, Transformers), containerizing training and inference pipeline for music generation in DeepComposer◦ Software development support for DeepLen and DeepComposer, including routine on-call tasks, e.g. updating Dashboard metrics, resolving SIM tickets from internal/external costumers, responding forum

◦ Build deep neural networks and pre-processing and post-pressing datasets for image denoising, inpainting.◦ Realize the latest convolutional neural networks (including UNet, FrameletsNets, Autoencoders, Deep Residual Learning, Generative Adversarial Networks (GANs), Reinforcement Learning) for medical imaging to improve disease detection and decrease the cost of the medical system◦ Support staff software engineers and research scientists in building prototype projects related to AI◦ Key points: Docker, PyTorch, Visdom, CNN, Wavelets, Nvidia DGX-1, PET/CT

◦ Wrote Python to examine and test the multifaceted performance of terrestrial ecosystem dynamics, with a total dataset of 30 years hourly data for 42 global sites and tenfold variables, e.g. carbon dioxide, water vapor◦ Exploited the relationships between observations and different Earth-Land Models of terrestrial ecosystem dynamics by arranging and adapting different statistical metrics◦ Key points: Python, Numpy, Matplotlib, Time Series