• Mathematically modeled the relationship between brain’s activity and blood flow using diagonal loading deconvolution and multi-variable optimization with stochastic gradient descent method; results were published in PNAS (2016).• Studied on the physical property of light attenuation and mathematical modeled optical interference with brain tissue with Monte Carlo simulation; results were published in Phil. Trans R. Soc. (2016), and widely adopted by the field.• Achieved functional brain region parcellation/dimension reduction with unsupervised classifiers, including k-means clustering, ICA, PCA, non-negative matrix factorization and canonical polyadic decomposition.• Extensive experience in image/data processing and exploratory data analysis with custom-written MATLAB scripts.• Designed workflow to process large data sets (100+ GB per sample) with parallel processing and GPU computing. • Developed and setup new paradigm for imaging awake animals, programmed the hardware in C to control physiological monitors, cameras and stimulators.

• Hands-on training in Python with a variety of data sets• Working in collaborative settings with Jupyter Notebook and Github• Participated in Capstone project in which designed and trained a Convolutional Neural Networks to classify MNIST dataset, achieved 99.5% accuracy. • Practiced on transfer learning by extracting variables from pre-trained VGG16 network and reapplied on new classification problem

• Studied the neurological mechanisms of brain recovery by analyzing electrophysiological signals with various processing methods including spatiotemporal filtering, trend removal and correlation analysis .• Produced statistical results and scientific figures with R and SAS.

• Implemented Artificial Neural Networks (multilayer perceptron and Recurrent Neural Networks) to decode the relationship between movement and signals from animal’s brain. Fine tuned the parameters and applied trained network to real-time brain-computer interface.