Triaged, analyzed, and fixed ambiguous performance regressions and bugs/crashes on levels from kernel, Android framework, up to feature/applications.• Daily tasks include: (a) designing and implementing device-interfacing automation tools to reproduce issues and collect data/logs, (b) debugging OS source code with patch builds, and (c) analyzing lab/field data and statistics. (python, bash script, Java, C++)• As STO of QA issue on development team, I own deep-diving all test errors. Enhanced every device performance tests to become more robust (10+ major bug fixes), take less time (50%+ reduction), and collect/report more accurate data (python/Pytest). Enhanced tuning procedures for 2 key OS parameters, which has been the standard procedures for 10+ new Fire TV/tablet products.• Developed novel tuning methodology to ZRAM configuration to optimize device performance, including high-level design, automation of experiment procedure and parameter sweeping, and development of data analysis algorithm based on Expectation–maximization algorithm (EM). It is the first data-driven procedure to tune this complex memory management component.• Built part of a key FireOS feature that improves app launch times. Developed full unit test suite of the feature using Mockito/Junit. Created robust cloud management script to prevent cloud storage cost from exponential growth while the feature is scaling up on more field devices. (AWS lambda, VPC networking)• Developed remote device lab which enable developing, configuring, and testing on our tablet/TV devices through web interface, video stream and command line. Developed cluster of multiple linux hosts with LXC containers, created a light-weight web interface with Flask, mySQL, webSSH.

•Using Ensemble Optimization Method (EOM) and other statistical and numerical methods to analyze Small Angle X-ray Scattering (SAXS) data, which is used to recover 3D structure of nucleosome core particle with methylated DNA.•Conduct molecular and physics experiments (SAXS, DNA extraction, transformation, etc.)

• Analyzed statistical data from annual credit rating quality report of Chinese bond market using Python codes. Created high quality model for default rate prediction • Write python codes to implement stochastic volatility implied (SVI) model Applied SVI on option pricing in Chinese market

• Genome data analysis: Compare different genotype of E. coli using online databases• Imaging data processing: Collaborate with postdocs to analyze images from microscope, extract features of E. coli growth using python.• Data modeling: using image data to model E. coli growth parameters with Matlab