Course Catalog & Credential Pathway
A sequence for students entering fields where AI does the drafting and a human is still the one who catches the error, frames the problem nobody templated, and signs their name to the outcome.
Five strands, running through every phase
No AI-specific coursework yet — it's too early to specialize, and the substrate underneath verification work is ordinary rigor: real statistics, real lab discipline, real domain choice.
Statistics & Probability I
Distributions, sampling, and why a plausible-looking result can still be wrong. Taught before calculus, not after — this is the branch of math the verification role uses daily.
Scientific Reasoning & Lab Practice
Designing a fair test, controlling variables, writing up a null result honestly. The habits, not the facts.
Argument & Evidence
Writing that has to survive someone disagreeing with it. Every claim cited, every claim defensible.
Computational Thinking
Decomposition, abstraction, debugging as a discipline — before any specific language or tool.
Domain Elective — Year 1
One real subject chosen for depth, not breadth: biology, structural engineering, civil law, whatever the student is drawn to. This choice matters more than any AI course on the page — it's the domain they'll eventually be qualified to verify.
AI enters the sequence here — deliberately late, and deliberately positioned as one applied practicum rather than the spine of the curriculum.
Experimental Design & Causal Inference
Confounders, A/B tests, why correlation in a benchmark isn't validity. Direct preparation for reading an evaluation and knowing whether to trust it.
MATH 101
Systems & Failure Modes
How software — and AI systems specifically — breaks: edge cases, distribution shift, confident wrong answers. Taught through case studies of real incidents, not lecture.
CS 100
Ethics & Decision-Making Under Uncertainty
Frameworks for a defensible judgment call when there's no rulebook — the actual daily work of the accountable-party role.
Domain Elective — Year 2
A second year in the same domain, now at a level where the student can identify a wrong answer, not just follow one.
ELECT 1
Applied AI Practicum
Working directly with current model-assisted tools inside the domain elective. The one course that's tool-specific — deliberately placed last, since tools rotate faster than everything above it.
CS 210
Not necessarily a four-year degree. Nursing program, engineering track, trade certification, paralegal pathway — the format follows the domain the student picked in Phase I, not the other way around.
Domain Specialization
Formal training in the chosen domain — this is the actual credential. Everything else in this track exists to support it, not replace it.
Applied Evaluation Design
Building and critiquing evaluations for AI-assisted output inside the student's own domain. The single most direct vocational skill in the sequence.
STAT 201
Professional Accountability & Documentation
How sign-off actually works in a regulated or high-stakes field: audit trails, liability, defensible documentation. Taught by practitioners, not theorists.
Supervised Practicum — Human-in-the-Loop
Placement inside a real workflow where AI output requires human verification. Graded on catches — errors actually found — not on output produced.
CAP 250, ACC 320
This phase has no end date and no credit total, on purpose. Every phase above is a line item; this one is a loop the graduate stays inside for the rest of their career — because the tools it's built around will be replaced repeatedly, even if the underlying role doesn't change.