Skip to content
View pj-workspace's full-sized avatar
🏠
Working
🏠
Working
1 follower · 0 following

Achievements

Achievement: QuickdrawAchievement: Pair Extraordinairex4Achievement: Pull Sharkx3Achievement: YOLO

Achievements

Achievement: QuickdrawAchievement: Pair Extraordinairex4Achievement: Pull Sharkx3Achievement: YOLO

Highlights

  • Pro

Block or report pj-workspace

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
pj-workspace/README.md

Pan Jie — From an idea to a working system

PJLAB portfolio GitHub Cursor profile Follow Jay on X Email Jay

I build at the intersection of AI, data, and product engineering—turning specific problems into systems people can actually use.

Current focus: agent workspaces, multimodal tools, and robotics data operations

CONTRIBUTION SIGNAL / PJLAB TRACE

PJLAB contribution trace generated from Jay's GitHub activity

About

I’m Pan Jie (Jay), a software engineering undergraduate and full-stack AI developer. My work spans agent workspaces, multimodal data systems, robotics operations, and native macOS tools.

What connects these projects is a bias toward complete systems: the interface, service, data model, and operational feedback should make sense together—not only in a demo, but throughout the real workflow.

  • Now: building agent workspaces, multimodal data tools, and robotics data infrastructure
  • Practice: full-stack product development, context engineering, interaction design, and system integration
  • Next: refining the systems already in motion and preparing for graduate study in artificial intelligence in 2027

Selected systems

Alcuin

AGENT WORKSPACE · PRE-ALPHA

A workspace for long-running agent conversations where context, knowledge, tools, traces, and approvals remain inspectable.

Next.js FastAPI LangGraph MCP Qdrant

View repository ↗

Multimodal Data Platform

PROFESSIONAL WORK · LIVE / EVOLVING

A collaborative annotation and review platform spanning images, audio, video, text, and point clouds. I hold sole development responsibility across import, workspace, review, and delivery flows.

Vue 3 TypeScript Spring Boot Point Clouds

Read the overview ↗

OpenRoboOps

ROBOTICS DATA OPERATIONS · OPEN SOURCE V0.1

One console for robot telemetry, camera previews, collection tasks, and datasets—with resumable transfers, integrity checks, and safety-conscious device adapters.

Next.js FastAPI PostgreSQL AsyncSSH WebSocket

View repository ↗

FlowShelf

NATIVE MACOS UTILITY · OPEN SOURCE V0.1

A lightweight floating shelf for moving files between apps while preserving Finder semantics, source locations, Quick Look, and the rhythm of the current task.

Swift 6 AppKit SwiftUI Quick Look

View repository ↗

Engineering stance

01 · Make context traceable.
Compression should retain a path back to its sources; tool execution should remain understandable after the page reconnects.

02 · Let interface certainty match system reality.
An interaction looking complete is not the same as its data being saved, reviewed, and ready to move forward.

03 · Respect the system around the tool.
The best small utilities fit existing workflows and platform semantics instead of asking users to relearn them.

Working set

Area Tools and technologies
Languages TypeScript · Python · Java · Swift
Product surfaces React · Next.js · Vue 3 · SwiftUI · AppKit
Services & data FastAPI · Spring Boot · PostgreSQL · Qdrant · WebSocket
AI systems LangGraph · Spring AI · RAG · MCP · OpenAPI
Delivery Docker · GitHub Actions · Nginx · Linux

More explorations

  • DID SAT — decentralized identity resolution and TLS security assessment
  • TCM Knowledge Agent — domain inquiry with retrieval and structured knowledge
  • ZhiXing — AI-assisted travel planning with Spring AI
  • Video Data Analysis — data collection, text analysis, and visual exploration

An idea worth building together?
AI products · open-source projects · problems that are still taking shape

pjlab.top · cursor · x · email · gitee

Popular repositories Loading

  1. tcm-intelligent-inquiry tcm-intelligent-inquiry Public

    An AI-powered Traditional Chinese Medicine (TCM) intelligent inquiry system for symptom analysis, constitution identification, and personalized health recommendations

    TypeScript 2

  2. vision-mcp-server vision-mcp-server Public

    Python

  3. qq-mail-mcp-server qq-mail-mcp-server Public

    Python

  4. plantuml-mcp-server plantuml-mcp-server Public

    Python

  5. DS-Fathom DS-Fathom Public

    macOS menu bar app that displays your DeepSeek API account balance in real-time with charts, notifications, and history

    Swift

  6. cursor-usage-menu-bar cursor-usage-menu-bar Public

    Native macOS menu bar app for Cursor usage, spending, and billing — SwiftUI dashboard with dual-pool analytics, charts, and EN/ZH i18n.

    Swift