Sean Kim

AI & CAD/CAM Software Development Portfolio.

I researched AI for recognizing machining features and furniture panels in 3D models, then used the recognition results to generate Fusion CAM operations and calculate furniture resizing.
Within the Wedraw Pro team, I led manufacturing-data output and production integration, connecting furniture design to production.

01

AI·CAD/CAM·BlueChain add-in

2025–2026

AI that reads 3D models and designs processing strategies

Fusion CAM automation: Geometry analysis → AI-based machining feature recognition → Machining strategy design → Visualization of digital twin results.

PythonOpenCascadeTensorFlowAutodesk Fusion APICAM

Assigned role

I personally researched and developed geometry analysis, machining strategy design and Fusion integration. In the digital twin project, my role was limited to visualizing simulation results. A handled the digital twin (MDT) simulation; I visualized the results it provided.

Result

Recognizing machining features in 3D models to automate Fusion CAM operations

  • 9 types Machining feature to identify · Include stock
  • ±0.15 mm Criteria for determining over-machining based on design tolerances (MDT example)
View Project
02

AI · Furniture · Parametric Design

2025–2026

AI that recognizes furniture panels and calculates resizing rules

Model analysis → AI-based feature recognition → Formulating parameter relationships → Generating and verifying manufacturing data

TensorFlowHierarchical CADNet basedGraph inputLinear regressionBOM/Drawing

Assigned role

I led the AI research in this team project. We defined the furniture rules together and jointly developed manufacturing data generation.

Result

When the dimensions of the furniture are changed, the panels adjust automatically according to predefined rules.

  • 15 types Panel types recognized by AI
  • 13 Design parameters defining individual pieces of furniture
View Project
03

Product · SketchUp · Furniture Manufacturing

2026 · In progress

Wedraw Pro, from furniture design to production

Designers create designs using 'Furniture Builder,' while manufacturing personnel receive production data via a connector.

SketchUp Ruby APIJavaScriptPythonMPR·PGINesting

Assigned role

This product is developed with a team. I led manufacturing data output and its introduction into production, as well as the initial development of drawings and ordering. The team jointly developed the furniture builder design tool.

Result

The designer's design is converted into machining data for the factory.

  • 4 types BOM, label, processing drawing, processing code output
  • 2 types Machining files tailored for nesting machines and the Skipper
View Project

Project Archive

Development work I took on beyond the key projects.

04

Operational Automation · Service Stability

2026

A monitoring system that detects issues spanning from the server to the user's PC.

Monitoring server and operational outcomes → Collecting API/PC app errors → Sending Slack alerts → Tracing root causes via the administrator console.

PythonFastAPIRubyReactTypeScriptPostgreSQLSlack API

Assigned role

Designed monitoring criteria and notification workflows. I developed everything from monitoring programs and server status dashboards to tools for verifying operational results, collecting API and PC application errors, and an administrator console to aggregate and view all this information.

Result

The system identifies operational failures that might otherwise go unnoticed for weeks—even when the server itself appears to be functioning normally.

  • 0.5 seconds → 0.1 seconds Average response time of the Order Service database, observed immediately after improving slow queries
  • 6 cases Operational issues identified and resolved via monitoring alerts
View Project
05

Data Engineering · Web Behavior Analytics

2021.06–2022.01

Redesigning the Architecture of VODA, a Web Behavior Analytics Service

Click and scroll collection → SQS message processing → Elasticsearch storage → periodic aggregation → heatmaps and charts

PythonSeleniumDockerElasticsearchMySQLAWS SQS

Assigned role

I fixed click-coordinate collection and built the processing that turns SQS messages into Elasticsearch documents, a refiner that prepares aggregates in advance, and a crawler that captures full pages.

Result

Pre-computing aggregates fixed the timeouts on the analytics screens

  • 5 cases Collection and aggregation problems I solved myself
  • 4 Components I built
View Project
06

Data Engineering · Recommendation Service Design

2021.03–2021.06

VORY-Q, a Personalized Curation Service for Online Stores

VODA page views and purchase history → customer segments → statistics-based recommendations → shown in the store through a header script

PythonDjangoMySQLCeleryRabbitMQAWS SQSPandas

Assigned role

As software architect, I designed data collection and analysis, how recommendations are displayed, the REST API, the relational database, customer segments and the recommendation algorithm structure, and I developed the statistics-based recommendation algorithms.

Result

I designed it so a store only adds a script to its header to show recommended products

  • 6 areas Designed by me, from the data flow to the recommendation algorithm structure
  • 3 Statistics-based recommendation algorithms I developed
View Project
07

Desktop App · Video Processing

2020.11–2020.12

A Video Player That Marks Scene Changes

Open a video → run scene splitting → mark split points on the seek bar → preview and jump between scenes → save a segment

C#WPFOpenCvSharpFFMEffmpeg

Assigned role

I developed every player feature apart from the scene change detection algorithm: split-point markers and previews on the seek bar, timeline zoom, scene navigation, editing and saving split points, saving segments with ffmpeg, and the media information window.

Result

Scene changes show up on the seek bar, one click away

  • 8 Functions in the detail panel
  • 2 things Scene detection methods wired into the player
View Project

Patents and Programs Derived from Shape Research

The 3D furniture modeling technology was jointly invented and applied for a patent.
A registration application has been submitted for the labeling software used in machining feature recognition.

PatentApplication · Co-invention

Apparatus, method, and recording medium for 3D furniture modeling using the analysis of topological connections between furniture components

This technology models 3D furniture by analyzing the connections between its constituent components. I was responsible for the research on AI-based panel recognition related to this invention.

User contribution

AI Research and Development Manager / Co-Inventor
  • Representation of geometric data and contact relationships between panels and surfaces
  • Research on AI for panel recognition based on hierarchical graphs
Application Date
2025.11.21
Application Number
10-2025-0177712
AI that recognizes furniture panels and calculates resizing rules
ProgramRegistration application received

Labeling Software for Training AI to Recognize Machined Shapes and Features

This is a tool that creates learning data for feature recognition AI by displaying machining features on each face of the CAD model.

User contribution

Overall Program R&D
  • Displaying face-based features and editing classifications.
  • Saving shapes and labels, and transmitting to the server
  • Geometry graph and training data composition
Creation Date
2025.05.13
Date of registration application receipt
2025.12.12
AI that reads 3D models and designs processing strategies