👋 Welcome to Pupba Lab

Exploring AI Pipelines, MLOps, and Open Source🐇✨

Hi! I'm Gwangwon Jung,
a Vision AI & Generative AI engineer.

My main areas of interest include:

• 🖼️ Image, Video generation and processing
• 🧠 Large Language Models (LLMs)
• 👁️ Computer Vision
• 🛠️ MLOps and pipeline automation

A modular, production-ready image generation pipeline inspired by ComfyUI but designed for MLOps, automation, and large-scale batch workflows.

• Pythonic, class-based pipeline definition (Executors, Tunnels, Tools)
• YAML config for reproducibility
• Easy extension and custom models
• Built for enterprise/research/automation

A Windows-adapted repo for running BAGEL-DFloat11 models with full CUDA 12.4 and RTX 4090 support.

• Windows support for DFloat11-based multimodal LLMs (e.g., BAGEL)
• Easy setup with all required wheels and install scripts
• For researchers and developers looking for fast, high-precision LLMs on Windows

An open tutorial built with LangChain and LangGraph, designed for hands-on learning via Jupyter Notebooks.
I contributed to the VectorStore module, including the base tutorial templates and integrations with Chroma and Qdrant.
More contributions are planned in the future!

A custom node for ComfyUI designed to support batch-based processing within workflows.
It enables users to iterate over input lists during generation and can be deployed on AWS, where it integrates with EventBridge to handle batch-completion or error notifications.

• 📧 Email: rhdk5148@gmail.com
• 🌐 Blog: https://pupbani.tistory.com/
• 💼 LinkedIn: https://www.linkedin.com/in/gwangwon-jung-5287a7275

Built with ❤️ by Gwangwon Jung | Powered by Open Source