> Initiating diagnostic...
> Scanning local environment...
> Detected Hub: ISU AI Workshop
> Status: Agentic Persona Engaged.
> Ready for deployment.
> _
Live Workforce Status
The Tooling Ecosystem
Autonomous units students will learn to pilot and orchestrate.
Gemini
Claude
OpenAI
Perplexity
Copilot
Jules
OpenCode
Ollama
Tavily
Exa
Mistral
Antigravity
MSS26 Live Schedule
April 7 - 10, 2026 | Interdisciplinary AI Week | All times CEST
| Time | Tuesday, Apr 7 | Wednesday, Apr 8 | Thursday, Apr 9 | Friday, Apr 10 |
|---|---|---|---|---|
| 10:00 - 11:00 |
Lecture
The Local Intelligence Shift
From chatbots to agents. Tokenization, context windows, and local LLMs.
|
Lecture
The Neural Link Terminal
IDE mastery, markdown files, AGENTS.md memory, and the terminal workflow.
|
Lecture
APIs, Python & Automation
API keys, .env files, Python scripts, MCP, and the research pipeline.
|
Lecture
Ship It: GitHub & Deployment
Git, GitHub Pages, Firebase, web app fundamentals, and the capstone brief.
|
| 11:00 - 12:00 |
Capstone
Web App Dev Sprint
Build your capstone web app. Space-themed portfolio piece for the showcase.
|
|||
| 12:00 - 14:00 | 🍽️ Lunch Break | |||
| 14:00 - 15:00 |
Build with AI
Your First Local Agent
AI Toolkit setup, run a 7B model on your hardware, and deploy a multi-tool prototype.
|
Build with AI
The 1-Hour App Sprint
From zero to deployed web app using only the terminal + AI agents.
|
Build with AI
The Autonomous Research Engine
Build a lit-review pipeline with Tavily, Exa, and Semantic Scholar APIs.
|
|
| 15:00 - 16:00 |
🏆 Demo Day
App Showcase & Voting
Present your web apps. Gold, Silver, Bronze + People's Choice awards.
|
|||
| 16:00 - 17:00 |
Office Hours
Open Lab & 1:1 Support
|
Office Hours
Open Lab & 1:1 Support
|
Office Hours
Open Lab & 1:1 Support
|
🎖️ Ceremony Awards & Closing Gold, Silver, Bronze presentations. People's Choice award. |
The Full Curriculum
Prerequisite: The Readiness Protocol
Before entering the 4-day sprint, ensure your "Digital Backpack" is packed. We'll move from a consumer environment (Browser) to a builder environment (IDE).
Local Intelligence
Breaking browser dependency. Running models locally and understanding the math of tokens and context.
The Neural Link Terminal
Learning the IDE, markdown files, persistent AI memory with AGENTS.md, and the terminal + chat development loop.
APIs, Python & Data
Connecting AI to external data. API keys, environment variables, Python automation, and the research pipeline.
Demo Day & Awards
Ship your capstone web app. Final arena sprint at 2pm, then app showcase, voting, and Gold/Silver/Bronze awards at 3pm.
Labs & Resources
The Computing Stack
Where AI fits: Data, Computing, Communications, Mobile, and Intelligence.
Module 0: Tokenization
The atomic unit of LLMs. Visualization and breakdown.
Module 0.1: Context Windows
Mastering In-Context Learning and "Sparks of AGI" experiments.
Lab: CLI Mastery
Hands-on project organization and terminal command missions.
Lab: NotebookLM
Grounded synthesis and audio overviews for research docs.
Lab: Local Engine
Running Llama/Phi-3 locally with Ollama and Fine-Tuning.
Lab: Phi Silica LoRA
Fine-tune Phi Silica with LoRA in AI Toolkit.
Lab: HF Fine-Tuning
Trainer API fine-tuning on Yelp reviews.
Lab: Multi-Tool Agent
Building a Plan-Search-Verify agent on HuggingFace.
Lab: Space App Build
Deploying a Folium/Skyfield orbital tracker in 1 hour.
Lab: Ghost Defense
Using an adversarial agent to critique your thesis logic.
Lab: Copilot Chat
Tables, Diagrams, and Synthesis for Space Research.
Terminal Guide (Win)
PowerShell mastery for agentic control.
Terminal Guide (Mac)
Zsh and Bash commands for AI automation.
Adversarial Primer
Theory and practice of agent-on-agent critique.
Integration Primer
Connecting agents via APIs, Opal, and MCP.
Agent Skills Guide
Architecting persistent instructions and .md skills.
Terminal AI Guide
Gemini CLI, Claude Code, and OS-Level Intelligence.
GitHub Copilot CLI
Agentic coding with Plan Mode and MCP integration.
Install Copilot CLI
Setup guide for WinGet, Homebrew, and npm.
AI Toolkit for VS Code
Fine-tune and run local models from VS Code.
Codex in VS Code
Install OpenAI’s coding agent extension.
Live Preview
Preview your site locally inside VS Code.
MCP Protocol Guide
The "USB-C" of AI. Connecting tools without code.
Model Comparison
GPT-5, Claude 4.5, and choosing the right brain.
Lab: Calendar Sync
Automate your agent to read and brief you on your daily schedule.
Readiness Protocol
Setup guide and mental shift for the workshop.
Lecture References
Supplemental readings and videos discussed during lectures.
Project Glasswing
Anthropic's $100M initiative uniting AWS, Apple, Google, Microsoft, and NVIDIA to use Claude Mythos Preview for defensive cybersecurity. The model found zero-day vulnerabilities in every major OS and browser, some hidden for 27 years. A landmark case study in AI-augmented security.
anthropic.comAI In Context
A focused YouTube channel covering AI concepts, tools, and workflows in practical, digestible episodes. Great for staying current with the rapidly evolving AI ecosystem and understanding how new models and techniques fit into real-world applications.
youtube.com/@AI_In_ContextThe Last 6 Decades of AI, and What Comes Next
Ray Kurzweil traces 61 years of exponential growth in computation (a 75 quadrillion-fold increase) and forecasts AGI by 2029, longevity escape velocity in the 2030s, and the Singularity by 2045. A 13-minute masterclass in exponential thinking.
TED · Ray Kurzweil · 13 minAllen Institute for AI (Ai2)
The nonprofit AI research institute founded by Paul Allen. Home to OLMo (truly open LLMs), AlphaFold-scale science tools, embodied AI for robotics, and planetary-scale environmental models. A model for open, mission-driven AI research.
allenai.orgAsta: Scholarly Research Assistant
Ai2's agentic research assistant powered by 108M+ abstracts and 12M+ full-text papers. Find papers, generate literature reports, and analyze data with natural language queries. A powerful alternative to manual literature reviews.
asta.allen.aiAgentic AI Foundation (AAIF)
A Linux Foundation initiative advancing open standards for agentic AI. Home to the Model Context Protocol (MCP), AGENTS.md specification, and Goose (open-source AI agent). The neutral ground where the agentic ecosystem is being standardized.
aaif.ioAGENTS.md
The open standard for giving AI coding agents project-specific instructions. Like a README for your AI teammates. Supported by OpenAI Codex, GitHub Copilot, Jules, Cursor, Gemini CLI, and more. Now stewarded by the AAIF under the Linux Foundation.
agents.mdModel Context Protocol (MCP)
The official documentation for MCP, the open-source standard for connecting AI apps to external systems. Think of it as the "USB-C port" of AI: one protocol to connect agents to databases, APIs, calendars, and tools everywhere.
modelcontextprotocol.ioGemma by Google DeepMind
Google's most capable open model family. From the 270M parameter edge model to Gemma 4's frontier reasoning, these models run on phones, laptops, and servers. Includes specialized variants: MedGemma (health), ShieldGemma (safety), and FunctionGemma (tool calling).
deepmind.google
Dr. Sounny-Slitine
Anwar Sounny-Slitine is an expert in GeoAI and a Faculty member at the International Space University. His research focuses on the intersection of Geography, Artificial Intelligence, and Spatial Data Science. He has architected numerous agentic systems for academic research and is passionate about bridging the gap between traditional social sciences and advanced autonomous computation.
Start Learning
Ready to move from AI consumer to AI architect? Build your own autonomous workforce today.