Get ready for an exciting journey where innovation meets cutting-edge technology! 🌟
This hackathon challenges you to create impactful tools for Red Hat Consulting using the Red Hat OpenShift AI portfolio: Workbenches, Data Science Pipelines, Model Serving, and more.
Your mission? Leverage Docling, the open-source document processing tool, and integrate it with Red Hat OpenShift AI to build intelligent, efficient document processing pipelines. The goal is to transform unstructured document data into actionable insights using the power of generative AI. 💡
With Red Hat OpenShift AI’s robust hybrid cloud platform ☁️, participants can build scalable, enterprise-ready pipelines for:
- Model Inference
- Fine-tuning generative AI models
- Retrieval-augmented generation (RAG) workflows
- Context-aware question-answering systems
This hackathon is the opportunity to explore these technologies and create impactful solutions for real-world problems.
Important
This main Readme refers to the starting point of the Hackathon. If you want to check the actual final solution, please, refer to this README file.
A hybrid cloud platform optimized for deploying generative AI workloads 🌐:
- Supports all levels of MLOps: develop, train, productize, and serve ⚡
- Combines IBM’s Granite LLMs with Red Hat’s InstructLab tools 🛠️
- Enables fine-tuning and deployment of custom models across cloud environments 🚉
An open-source tool for advanced document parsing and conversion:
- Supports formats like PDFs, DOCX, PPTX, and HTML 📑
- Integrates with tools like LlamaIndex and LangChain for RAG tasks 🦙
- Preserves context in complex layouts (e.g., multi-column text or tables spanning pages) 📊
An open-source vector database designed for similarity search and RAG:
- Manages large-scale embeddings from unstructured data sources 🔍
- Integrates with tools like LangChain for seamless retrieval workflows 🛠️
- Optimized for high-performance similarity search, enabling fast query results even with billions of vectors ⚡
- Supports hybrid storage (disk and memory) to balance cost and performance 💼
A versatile, self-hosted AI interface designed for maximum adaptability and security:
- Fully offline operation ensures data privacy and control 🔐
- Customizable workflows tailored to diverse use cases, from research to production 🔄
- Modular architecture supports seamless integration with third-party tools and APIs 🔗
- Intuitive interface simplifies interaction with advanced AI systems, enhancing productivity 🚀
This repo contains the following key folders:
examples: Base custom resource files, data science pipelines, and Python scripts used to build the project.project: The project implemented during the hackathon.docs: Documentation supporting the project implementation.
- André Lizardo | [email protected]
- Alvaro Lopez | [email protected]
- Nacho Lago | [email protected]
- Andreas Nixel | [email protected]
- Suresh Vishnoi | [email protected]
- Adam Bassett | [email protected]
- Abinash Ramesh | [email protected]
- Jack Hawkins | [email protected]
- Familiarize yourself with Docling by exploring its documentation 📘.
- Learn about Model Serving and Model Inference on OpenShift AI.
- Explore deploying AI workloads on OpenShift AI using Data Science Pipelines ⚙️.