Junho Kong - Developer Experience Portfolio

This is a personal portfolio site for full-stack manufacturing AI products, developer experience, and AI builder work.

Junho is a Manufacturing AI Platform Engineer at SK On, Codex Ambassador, and PseudoLab-grown builder. His work connects graduate manufacturing AI research, On-Platform, agent evaluation, Korean Codex onboarding, and community programs that help developers turn frontier AI into practical workflows.

The interface is structured as a Codex-inspired thread workspace so interviewers can scan Junho's projects by topic.

Live site: https://generoustandard.github.io/Portfolio/

Career Narrative

• Graduate manufacturing AI research: experimental data, image and time-series processing, statistical validation, machine learning, and deep learning.
• On-Platform at SK On: reusable full-stack analysis workflows for non-developer manufacturing users.
• Science Agent Leaderboard: correctness, tool usage, cost, execution validation, and reproducibility.
• Codex Community Korea: Korean-first developer onboarding, learning surfaces, and community programs.

Research Foundation

• M.S. in Mechanical Engineering, Sungkyunkwan University, 2020–2022; GPA 4.0/4.5.
• Developed a Directed Energy Deposition (DED) real-time anomaly-monitoring workflow using image preprocessing, STFT/HHT, and 2D-CNN/M2D-CNN models.
• Reduced the analysis cycle from one week to one day; the HHT-based approach improved model accuracy by approximately 20%.
• Built a DED multi-defect diagnosis pipeline using 40 image features, ANOVA and t-tests, and ANN/KNN/SVM classifiers; the best SVM model achieved 92.7% accuracy.
• This research foundation informs On-Platform's emphasis on validated analysis, evidence, traceability, and reusable workflows.

Focus

• Full-stack manufacturing AI platform engineering at SK On
• Korean Codex community onboarding and programs
• Agent evaluation and benchmark design
• Manufacturing AI research and experimental validation
• Creative AI and user-centered AI workflows
• Documentation quality and localization experiments

Featured Projects

• On-Platform
• Codex Community Korea
• Science Agent Leaderboard
• Codex-Community-Docs-KR
• Bon

Notes

This is a personal portfolio. Codex Community Korea is an ambassador-led community initiative and is not an official OpenAI website. On-Platform represents company work at SK On, described using sanitized synthetic scenarios without confidential manufacturing data.

Development

Install dependencies:

npm install

Run the local development server:

npm run dev

Build the production site:

npm run build

Preview the production build:

npm run preview