Mag.AI-Science — Program Overview
Castalia Institute
One sentence¶
Mag.AI-S is the science track of the Castalia Magisterium: you build testable world models, honest uncertainty, and reproducible lineage—then deploy with evidence, not hope.
Who it is for¶
Researchers who are tired of “AI for slides”
Lab and platform leads who want mechanisms, not vanity metrics
Builders who will use AI as instrument, not oracle
Where you learn: iNQspace¶
Mag.AI-S teaches AI through iNQspace — the Castalia lab environment where Jupyter notebooks execute, simulations run, tools connect under MCP policy, and traceable artifact lineage is recorded for faculty review.
How program materials are published: Jupyter Book + executable articles¶
Syllabi, lectures, and Jupyter notebooks are maintained as MyST / Jupyter Book source in this repository and published as static program sites on Castalia Magisterium—one Jupyter Book per course plus this program hub. That is the public, citable map of the curriculum.
Executable articles are the default instructional unit: narrative and code live in the same lineage—notebooks under each course’s notebooks/ (built into the course book) are developed and run in iNQspace, not as an afterthought PDF. Full policy: docs/EXECUTABLE_ARTICLES.md.
Scientific tools (MCP): Where the curriculum requires it, you leverage and build Model Context Protocol integrations so AI workflows connect to instruments, literature, lab systems, and data pipelines—with explicit scopes. Details: docs/MCP_SCIENCE.md, docs/CURRICULUM_FULL.md.
AI in the curriculum: The program embeds a technical AI literacy spine (evaluation, RAG/grounding, generative workflows, scoped tools). See docs/CURRICULUM_FULL.md (AI literacy spine).
What you build (not what you memorize)¶
Executable hypotheses you can perturb and falsify
Measurement and uncertainty encoded as first-class objects
Reproducible workflows with documented provenance
Governed automation with safety and integrity boundaries
MCP-linked toolchains for instruments and knowledge systems (where courses require it)
Credential¶
Magister of Artificial Intelligence in Science (Mag.AI-S) — earned through artifacts, defenses, and a final thesis deployment. See certificate.md.
First course¶
AINS-S6001 — Scientific World Models & Executable Hypotheses is the recommended entry: the smallest honest scientific world—assumptions, observables, and tests—before larger systems. Open the course book: AINS-S6001 (full list on Course sites).
Legal¶
The Castalia Institute Magisterium confers proprietary credentials based on demonstrated work and evaluation. These credentials are not accredited academic degrees and do not confer professional licensure.