What is AI VERDE?
AI, Virtual, Explorer for Research, Discovery and Education (AI Verde) is an open-source, in-house AI platform developed by the Data Science Institute at the University of Arizona. It connects researchers and educators to leading AI models through a unified interface for building chatbots, running custom workflows, and integrating GenAI into research and teaching. With built-in privacy, budget controls, and institutional data protection, AI VERDE empowers users to explore and deploy AI solutions confidently.
Unique Solution
AI-Verde’s unique feature is its unified access to multiple large language models (LLMs), including commercial providers like GPT-4, Claude, and Gemini, as well as open-source and on-premise models like LLaMA4, Gemma, and PHI-4—all within a single, managed platform. This centralized access allows users to test, compare, and fine-tune models in one place, streamlining workflows and integrating institution-specific data securely.
AI-Verde enables faster experimentation, reproducible workflows, and precise budget control. It also provide a secure, classroom-ready environment with customizable permissions, reusable prompt libraries, and temporary access options tailored for workshops and academic courses.
- In-House Privacy – Data processing occurs within the university's infrastructure, safeguarding information.
- Train Specific AI – The development of AI models tailored to institution-specific data, enhancing accuracy.
- User-Controlled Access – Enables admins to manage permissions and monitor AI use.
- Course Integrations: Rapid creation by converting existing training materials into engaging content.
- Reusable Prompts – Allows users to build a library of prompts, streamlining AI interactions and consistency.
- High-Performance Hardware – Provides access to cutting-edge GPUs and storage through Jetstream2.
Currently, AI Verde is available only to the University of Arizona community and a U of A NetID is required. If you are U of A faculty, staff, or researcher and want to use AI Verde for a course or research, email mithunpaul@arizona.edu. We will respond and set up an initial consultation to identify your specific needs and discuss how the AI Verde team can help.
Students who want access to AI-Verde, email Mithun Paul at mithunpaul@arizona.edu.
Yes! We offer training and consultations to help with your unique circumstance. Our team of AI experts are here to help.
- Students, faculty, and researchers who want to use AI Verde as a chatbot to ask questions to the latest LLM, sign up for training to get started.
- For those who want to learn what AI is, how AI works, or how to use complex AI tools provided by AI Verde through programming, we offer training workshops throughout the academic year at the U of A DataLab. Learn more about the various workshops and which ones are right for you.
- If you are not sure how AI can be used to solve some of your research problems, email mithunpaul@arizona.edu to reach the AI Verde team. Our AI experts can consult with you to determine how AI Verde can help.
- When your department wants to learn how AI can be leveraged, our AI team of experts can provide consultation. Email mithunpaul@arizona.edu to reach the AI Verde team.
- For faculty who want to create a chatbot trained on their course material, contact Mithun Paul at mithunpaul@arizona.edu and to show you how AI Verde can help.
For questions from U of A faculty, staff, or researcher for using AI Verde for a course or research and students who want access to AI-Verde, email Mithun Paul at mithunpaul@arizona.edu.
Yes. Learn how AI Verde can support academia and research. Find a comprehensive list of all AI Verde features in the research paper, "AI-VERDE : A Gateway for Egalitarian Access to Large Language, Model-Based Resources for Educational Institutions" by Paul Mithun, Enrique Noriega-Atala, Nirav Merchant, and Edwin Skidmore.
When using AI Verde for your research, please use this
BibTex citation: @article{mithun2025ai, title={AI-VERDE: A Gateway for Egalitarian Access to Large Language Model-Based Resources For Educational Institutions}, author={Mithun, Paul and Noriega-Atala, Enrique and Merchant, Nirav and Skidmore, Edwin}, journal={arXiv preprint arXiv:2502.09651}, year={2025} }