Sep 2023 – Present · 2 yrs 8 mos · Tokyo, Japan · On-site

• Full stack engineering role at pre-seed startup.

Aug 2022 – Mar 2023 · 7 mos · Seattle · Hybrid

• Architected full-stack computer vision Android application for guiding the installation of edge computing devices throughout Amazon fulfillment centers.
• Created a camera-based target tracking system for said application that determines target focus level while also rendering an options and information user interface, optimized to run at a full, stable 30 FPS.
• Built and iterated upon an accessible user interface for non-technical camera installers, decreasing the ~10% error rate to nearly 0% for vision-based applications in warehouses stemming from poor installation and speeding up the process by ~3x.
• Developed automatic exposure and gain algorithm to enhance readability of warehouse targets in extreme light conditions.

Jun 2021 – Sep 2021 · 3 mos · Seattle, Washington, United States · Remote

• Designed and implemented a computer vision Android application. Tracks multiple objects’ position, and movement while supporting objects either splitting apart or fusing within view.
• Increased robustness and accuracy by iterating through and solving common computer vision weaknesses.
• Created custom API to interact with internal data for internal consuming applications.

Nov 2020 – Jun 2021 · 7 mos · San Diego, California, United States · Remote

• Lead the first steps towards Jira issue automation based on security and vulnerability alerts found during static code analysis from OWASP Dependency-Check and SonarQube.
• Refactored all Jenkins shared library files.
• Implemented complete unit test coverage using Jenkins Spock.
• Standardized Jenkinsfiles across internal repositories.

Jun 2020 – Sep 2020 · 3 mos · Irvine, California, United States · Remote

• Implemented sequential pattern mining to repurpose previously collected data into more readable data designed for non-technical clients’ understanding.
• Revised the PrefixSpan algorithm to work 60x faster with The Trade Desk’s data storage format while keeping memory costs extremely practical.
• Refurbished a client-facing webpage and added prompting to show newly mined data.
• Product is planned to go public for The Trade Desk’s 2020 Winter UI Release. Backend is currently running daily.
• Worked extensively in Spark, Scala, AWS EMR, Atlassian Stack, Apache Airflow, C#, and React.