James Park

Software Development Engineer

Last updated: 23 Feb 2026 Download resume PDF

Currently based in Seattle, WA

Designed and delivered high-scale, security-first features for AWS Systems Manager, serving a customer base of 100,000+ and leveraging Java, Golang, TypeScript, and AWS services (S3, CloudFormation, IAM, DynamoDB, EC2, CloudWatch, SQS, SNS).
Led a 14-month project to build an AWS SSM microservice in 7 isolated ISO-certified regions, requiring close coordination with multiple engineers with ISO access and bringing AWS services in ISO regions up to parity with commercial regions.
Built an additional UI component in the service's AWS web console for accessing the SSM Documents API (AWS SDK, TypeScript, React, Jest, Sinon).
Optimized the AWS SSM Session Manager microservice's ability to stream to AWS CloudWatch Logs in real time (Golang).
Reduced the AWS SSM Health microservice’s region expansion time from 2 months to 2 weeks by creating comprehensive, standardized documentation that streamlined the process for 3 new regions.
Transitioned the service's region expansion efforts from a manual process to an automated zero-touch approach, saving ~100 man-hours per new region (CloudFormation, IAM, internal CI/CD tools).

Relevant Coursework: Object Software Design (JS), Computer Algorithms (Python), Artificial Intelligence (Python)

3.5/4 GPA, Dean's List
Relevant Coursework: Computer Science I+II (C++), Data Structures, Info and Data Management (Python, SQL), Intro to Video Game Design (C#), Creative Coding (JS)
Member of Design for America, an idea incubator tackling social challenges through design

Replaced COLMAP-based Structure-from-Motion step from previous iteration with GLOMAP (Global SfM) and LightGlue (Neural Feature Matching) to increase pose accuracy for challenging dataset.
Integrated Physically-Plausible ISP (NVIDIA PPISP) into the training loop to dynamically correct auto-exposure and white-balance drift across large-scale image dataset.

Architected a cloud-native 3D reconstruction pipeline using Nerfstudio on AWS G5 instances to transform a 1.2k image unstructured dataset into a high-fidelity Gaussian Splat of a large mall interior.
Integrated AWS S3 Mountpoint for data staging, enabling real-time streaming of multi-GB datasets to the training environment and reducing local storage overhead by 90%.
Ported the digital twin into an interactive first-person walkthrough in Unreal Engine 5, using Luma AI plugin for real-time rendering of Gaussian splats through Niagara.

Moderating and troubleshooting Mozilla Hubs (A-Frame) social spaces designed for presenting postgrad CS research focused on graphics, UX research for VR

Utilized Django (Python), SQLite database, and Bootstrap CSS to design and develop a web application that allows users to search for books they want to borrow from other users in their area.

Utilized Unity SoundMixer to create a dynamic audio experience, implement 3D positional audio, and manage and adjust audio settings in real-time.
Source control via git (Bitbucket) and tracking via JIRA and Unity collaboration features to work with a team of developers on a large project.
Managed project completion on a tight 3-day schedule for a professionally evaluated product.
Developed and maintained productive working relationships with team members involved in industry software engineering and 3D asset creation.

Demonstrated features in Magic Leap One LuminOS, Oculus Go, and SteamVR ecosystem to 1000+ industry professionals and researchers.
Tested a virtual reality application with an AI conversational agent designed to react dynamically to voice input.
Operated specialized VR equipment for accessibility (eg. voice commands, custom controllers, eye tracking).
Demonstrated a strong commitment to generating interest among colleagues in bleeding-edge mixed reality hardware.