Official website for the AI Innovation Team at Red Hat and IBM, showcasing our open-source AI research, tools, and publications.
🌐 Live Site: https://ai-innovation.team
We are an AI research team at Red Hat and IBM with collaborators from MIT, UMass, and other leading institutions, dedicated to fostering an inclusive and accessible open-source AI community. Our mission is to empower individuals to collaborate in advancing the future of LLMs and generative models.
- its_hub - Inference-time scaling for LLMs
- sdg_hub - Synthetic data generation pipelines
- training_hub - Post-training algorithms for LLMs
- async-grpo - Asynchronous GRPO for scalable reinforcement learning
- hopscotch - A method for skipping redundant attention blocks in language models
- mini_trainer - Efficient training library for LLMs up to 70B parameters on a single node
- orthogonal-subspace-learning - Adaptive SVD-based continual learning
- probabilistic-inference-scaling - Inference-time scaling with particle filtering
- reward_hub - State-of-the-art reward models
- SQuat - KV cache quantization for scaling inference time
- training - Efficient messages-format SFT library
This website is built with Jekyll and hosted on GitHub Pages.
- Ruby (version 2.7 or higher)
- Bundler
- Jekyll
- Clone the repository:
git clone https://github.com/Red-Hat-AI-Innovation-Team/Red-Hat-AI-Innovation-Team.github.io.git
cd Red-Hat-AI-Innovation-Team.github.io- Install dependencies:
bundle install- Run the development server:
bundle exec jekyll serve- Open your browser to
http://localhost:4000
.
├── _config.yml # Site configuration
├── _data/
│ ├── settings.yml # Site settings, menus, projects
│ └── videos.yml # Video content catalog
├── _includes/ # Reusable HTML components
├── _layouts/ # Page layouts
├── _pages/ # Static pages (About, Publications, etc.)
├── _posts/ # Blog posts
├── _sass/ # Stylesheets
├── images/ # Image assets
├── js/ # JavaScript files
└── index.html # Homepage
Create a new file in _posts/ with the format YYYY-MM-DD-title.md:
---
layout: post
title: Your Post Title
date: YYYY-MM-DD
image: '/images/posts/your-image.jpg'
tags: [tag1, tag2]
---
Your content here...Edit _pages/publications.md and add a new publication card:
<div class="publication-card">
<h3><a href="ARXIV_URL" target="_blank">Paper Title</a></h3>
<strong>Authors:</strong> Author 1, Author 2<br>
<details>
<summary>View Abstract</summary>
Abstract text here...
</details>
<strong><a href="ARXIV_URL">📄 Arxiv</a> | <a href="GITHUB_URL">💻 Code</a></strong><br>
<strong>Date:</strong> YYYY-MM-DD
</div>Edit _data/videos.yml:
- title: "Talk Title"
speaker: "Speaker Name"
youtube_id: "VIDEO_ID"
youtube_url: "https://www.youtube.com/watch?v=VIDEO_ID"
embed_url: "https://www.youtube.com/embed/VIDEO_ID"
date: "YYYY-MM-DD"
category: "Random Samples"Edit _pages/about.md and add to the appropriate section (Team Members or Collaborators):
<div class="col col-3 col-d-6 col-t-12">
<div>
<a href="/images/about/Headshots/name.jpg" class="portfolio__link glightbox"
data-glightbox='title: Full Name; description: ; descPosition: bottom;'>
<span class="portfolio__icon"><i class="ion ion-ios-eye"></i></span>
<img src="/images/about/Headshots/name.jpg" data-src="/images/about/Headshots/name.jpg"
class="portfolio__image lazy" alt="Picture of Full Name">
</a>
<a href="PERSONAL_URL" target="_blank">Full Name</a>
</div>
</div>- ✅ Responsive design with dark/light mode
- ✅ Blog with pagination
- ✅ Publication showcase with expandable abstracts
- ✅ Video catalog with categories
- ✅ Project portfolio
- ✅ Team member profiles
- ✅ Google Analytics integration
- ✅ SEO optimized
- ✅ Fast performance with lazy loading
- ✅ Math support via KaTeX
This site uses a customized version of the Clancy Jekyll theme, optimized for research team presentation.
Content on this site is the property of the AI Innovation Team and respective authors. Check individual repositories for software licenses.
For questions or collaboration opportunities, please reach out through our GitHub organization.
Maintained by the AI Innovation Team | Red Hat & IBM | ai-innovation.team