Li Yu

Software Engineer at Discord

Berkeley, California, United States

About

Backend Software Engineer. Interested in algorithms, data analysis, and machine learning.

Experience

  • Software Engineer at Discord
    Jul 2021 - Present · 5 yrs

  • Backend Software Engineer at Sentropy Technologies
    Apr 2020 - Jul 2021 · 1 yr 4 mos

  • Software Engineer II at Narvar
    Sep 2018 - Apr 2020 · 1 yr 8 mos

    • Created the APIs and backend logic for the “Quin” language translation service, consolidating company-wide language translation work into one service and creating a single source of truth. Consolidation of language translation across the company improved language consistency across all products and enabled them to efficiently reach international markets, ultimately streamlining product launches. • Designed and implemented the database, codebase, and APIs for Narvar Concierge, a platform that enabled brick and mortar stores to become shipping and returns locations for online orders (featured in the WSJ). This allowed retailers to simultaneously expand their business models while increasing foot traffic to their stores.

  • Software Engineer at Keysight Technologies
    Jul 2017 - Sep 2018 · 1 yr 3 mos

    • Designed and implemented the interactive routing add-on for the company’s flagship product, allowing customers to stay within Keysight’s software rather than use a third party tool for routing. The improved functionality introduced by this feature helped streamline workflow for existing users and increased product adoption among potential clients. • Represented Keysight Technologies in Si2, a coalition of 69 industry-leading companies, research centers, and academic institutions in the electronics and EDA tool industries. As a board member of Si2, I have assisted in the creation of the collaborative software maintained by this coalition and ensure the needs of Keysight are addressed within the coalition.

  • Researcher at UC Berkeley Biomimetic Millisystems Lab
    Nov 2013 - Dec 2016 · 3 yrs 2 mos

    • Used real world sensor data of a robot to create an approximate simulation model of said robot, allowing the lab to run simulation tests that were less expensive and faster than physical tests. This filtered out experiments that did not show enough promise to be tested on physical robots, thereby reducing the cost and time of producing additional robots and running tests. • Wrote decentralized communication, landing, and object avoidance software for corporative aerial drones. This prototype was a precursor to a system of miniature drones to be used in search and rescue operations, allowing for greater area of coverage due to the number and size of the drones while also reducing the risk to emergency responders.