Student at University of Chicago

Algorithm Engineer Intern June 2018-Septermber 2018

Conducted algorithms for key point detection on a hand and bone age prediction, based on deep learning combing Resnet50, FPN and focal loss, using Mxnet framework and achieved 97% accuracy
Modified the original auto testing algorithm in lung nodules comparing process to achieve better generalization and higher efficiency with 50% less time. This feature was wrapped up in company’s source package
Designed and implemented a new interface for CT (computed tomography) scan results, according to the requirements from hospital, and long-term negotiation with hospital representative

Quantitative Analyst Intern May 2017-August 2017

Crawled future price data using SQL, processed missing data and conclude data reports bout price jumps, price limits and price spread
Conducted high frequency momentum strategy and implemented back-test on industrial data
Constructed a database containing all future contracts within latest 3 months’ prices and analyzed price changes between two nearby contracts
Reviewed high frequency trading books, investigated and presented weekly about trading strategies and discussed and brainstormed with the team

Research Assistant September 2016-December 2016

Generated new relocating data through latent variable and applied meta-analysis to combine result
Compared expected results with traditional mantel-haenszel method
Made theoretical identification of the superiority of the new method based on effienecy loss
Extended it to other scientific field, including data clustering, especially ordered data