John Jin

Experience

Education, work, research, teaching, and selected independent projects.

Education

September 2023 — March 2027 (expected)

University of Washington

B.S. in Computer Science · Minor in Informatics

Work

June — September 2026

Software Engineer Intern

Ignite Robotics

  • SummerTrack — Expanded an existing login and administration framework into a counseling platform using Next.js, TypeScript, Supabase, and PostgreSQL. Added institute administration and parent–student collaboration, with row-level security, revocable audience-specific sharing, and organization-scoped access controls.
  • Developed a five-stage, resumable pipeline to discover, extract, enrich, merge, and validate data for a 2,930-record catalog. With cross-source deadline resolution, explicit-date coverage rose from approximately 28% to 41.8% between scrape rounds, excluding rolling admissions.
  • Parts Request Portal — Contributed the catalog functionality of a Django parts platform: VEX catalog synchronization, SKU-level product variants, preservation of custom items during imports, and kit-aware search that surfaces containing kits for component-name queries.

June 2024 — September 2025

Software Engineer Intern

Trustrobot Co., Ltd.

  • Migrated HedraCAM’s OCCTProxy geometry and model data interfaces from C++/CLI to C# and native C++, using P/Invoke and pointer operations. Validated the interfaces on Linux and delivered them for production use.
  • Prototyped Python integration with application interfaces using ZeroMQ publish–subscribe messaging.

Research & teaching

September 2026 — Present

Research Room Program Co-Lead

UW Interactive Intelligence (I2)

  • Co-lead preparations for a research program expected to serve approximately 20 participants, organizing industry-speaker talks for outreach and developing a schedule of project review milestones.

Term expected to run through June 2027; program activities begin in Winter 2027. Providing research resources and publication support are planned responsibilities.

January — May 2026

Fellows Teaching Assistant

UW Interactive Intelligence (I2)

  • Mentored a team investigating LLM streaming, coordinating project work, reviewing milestone deliverables, and answering research questions while guiding topic selection and preparation for the final presentation and defense.

January — March 2025

Research Room Team Lead

UW Interactive Intelligence (I2)

  • Led a five-member team in analyzing multimodal architectures, defining the research scope through preliminary literature review, dividing paper reviews by subfield, and consolidating findings into a final presentation.

Research Room member from September 2024 to April 2025; team lead from January to March 2025.

Selected projects

Independent development

Repository

Broken Arrow Deck Builder & Data Toolchain

React · TypeScript · Python · Playwright

  • Independently developed and published a browser-based game deck editor on GitHub Pages, implementing unit availability, loadout pricing, transport compatibility, and budget validation in TypeScript.
  • Built a Python extraction and validation pipeline for 24 game-data tables containing 16,687 records, supporting card rendering and deck configuration for 540 units.
  • Implemented WebCrypto-based deck-file import/export and automated data-consistency and serialization round-trip checks, including Playwright workflows for UI-level verification.

One toolchain spanning BA-DeckEditor, BA-ReCard, and BA-Units. Game data and assets belong to their respective creators.

Independent development

November 2025 — Present

Repository

VLoid — Local Voice AI Desktop Application

Electron · FastAPI · SQLAlchemy · SQLite · SSE · WebSocket

  • Built a local voice AI desktop application with Electron and FastAPI, integrating local LLM inference, speech recognition, and text-to-speech through SSE and WebSocket interfaces.
  • Implemented persistent, branching conversations with SQLAlchemy and SQLite, supporting message history, active conversation branches, and configurable personas.
  • Developed a voice-input pipeline with voice activity detection, audio pre-roll, and partial/final transcription, alongside streaming speech synthesis with cached voice prompts in the voice-clone path.