• Created SafeBeauty, a webapp to detect unsafe skin care products using computer version.• Scraped 2000+ images of 36 skin care products with Scrapy and Beautiful Soup in Python.• Preprocessed the images with data augmentation and retrained an object detection model with tensorflow API using Inception v2 feature extractor and Faster R-CNN. • Achieved recall value 0.88 in validation and deployed the webapp with Flask on AWS platform.

• Analyzed 1 TB+ electrical and optical neural data to map stimuli with single neuron responses. • Built a multicompartment biophysical model to simulate single neuron responses and performed automatic optimization via evolutionary algorithm with 23 parameters.• Constructed a large scale (1000 neurons) network by incorporating the single neuron model.• Used parallel computing to run the network simulation on high performance cluster (HPC) and observed the predicted brain states transition emerging from them.• Boosted the efficiency of experimental design by 80% with computer simulation.

• Advanced the understanding of neural circuits changes during learning new motor skills.• Designed a toolbox in MATLAB to automatically track and analyze animal reaching behavior.• Explored memory related molecule changes using confocal image processing techniques, eg. automatic region of interest detection, 3D reconstruction, and adaptive thresholding.• Developed an autonomously updating Brain Machine Interface (BMI) using a reinforcement learning architecture for the control of prosthetic limbs.

Developed the neuroscience curriculum for 6th-8th graders

Teaching Assistant of the courses: Introduction to Cosmology, Introduction to MRI