Large data systems, machine learning systems.

Worked on multi-tenant data storage layer in C++, data pipeline engine in Python, and frontend data representation layer in Typescript/Javascript.

• Fixed defects that caused product crash (core dump); debugged C++ memory corruption issues.

• Investigated repeatability issue; dug into how parallel sorting of floating point numbers can be stable.

• Participated in war rooms to win million dollar contracts; helped to reduce leakage power by 4%.

• Built debugging tool into the product; significantly increased efficiency of analysis and debugging.

• Built an analysis command into the product; helped improve Quality of Results by 2% on average.

• Wrote analysis tools using Python scripts; modeled circuit data by bin-based logistic regression.

• Improved the product’s C++ code quality by refactoring issues such as memory leak, deadcode, etc.

• Developed a parallel string search engine with 2x performance on 4-core machine using 8 threads.

• Implemented Boyer-Moore search algorithm which runs 2% on average faster than KMP algorithm.

• Built a parallel model with producers that reads in the text and consumers that does the searching.