Seyeon Park

Resolved a rendering race condition in a chat UI's auto-scroll behavior using requestAnimationFrame and the scrollIntoView DOM API, preventing a pre-launch production issue affecting 100% of chat interactions and ensuring students and teachers always saw the most recent message without manual scrolling. Designed and built a Chrome MV3 extension prototype end-to-end (React, TypeScript, Vite/CRXJS) integrated with NextStep's internal AI grading gateway via a secure backend-for-frontend proxy, demonstrating core platform grading features to the engineering team. Rebuilt the extension's backend independently as a self-hosted Express API integrating Google Gemini and a Canvas LMS data pipeline, validating the grading workflow through user testing with teachers and graduate students.

Software Engineering Intern - NextStep AI Inc. - Seattle, WA

(2026-03 - 2026-06)

• Resolved a rendering race condition in a chat UI's auto-scroll behavior using requestAnimationFrame and the scrollIntoView DOM API, preventing a pre-launch production issue affecting 100% of chat interactions and ensuring students and teachers always saw the most recent message without manual scrolling.
• Designed and built a Chrome MV3 extension prototype end-to-end (React, TypeScript, Vite/CRXJS) integrated with NextStep's internal AI grading gateway via a secure backend-for-frontend proxy, demonstrating core platform grading features to the engineering team.
• Rebuilt the extension's backend independently as a self-hosted Express API integrating Google Gemini and a Canvas LMS data pipeline, validating the grading workflow through user testing with teachers and graduate students.

Technical Research Intern - Digital Scholar Program, School of IAS Lab - Seattle, WA

(2025-07 - 2025-08)

• Restored a non-functional flow injection analyzer (QuikChem 8500) to process 390 samples per run, troubleshooting complex hardware, fluidics, and software method configurations.
• Developed automated CSV data workflows and authored technical documentation for chemical protocols, eliminating false injections and standardizing data output for future operators.

Machine Learning Research Assistant - DAIS Lab - Seattle, WA

(2025-05 - 2025-11)

• Contributing to DeepTracer 3.0, generative AI model for protein structure prediction on multi-GPU infrastructure.
• Evaluated diffusion-based protein structure prediction pipelines, analyzing generative sampling outputs and model accuracy, achieving 1.74˚A RMSD and 75% coverage.
• Implemented HDF5 compression reducing dataset size by 20× and improving training speed by 14%.

B.S - Computer Science and Software Engineering, Minors in Data Science and Music - University of Washington (2026-08)