Bachelor of Science, AI Engineering course guide

Program shape and the catalog boundary

WGU lists 121 competency units for this bachelors program. The grid begins with E077, E066, E067, C458 and closes with E089, E090. It is a public standard path, not an individual transfer evaluation: accepted credit, licensure, prior learning, substitutions, program revisions, state requirements, and mentor planning can change a student's actual Degree Plan.

Technology plans braid two workloads: performance assessments that ship a working artifact with its write-up, and proctored objective exams over tools and theory. Keep code, configurations, diagrams, and reports on one evidence ledger so the capstone does not contradict design decisions made in earlier courses, and let the current Course of Study — not the course title — decide which instrument each class actually uses.

The catalog establishes membership and sequence, but it does not publish each course's assessment type or Task 1/Task 2 identities. Open the current Course of Study before planning deliverables. A course page on this site explains the reasoning method and program connections; a separate assessment manual appears only where WGU has publicly verified the real task.

The classes, one by one

How to read a WGU standard path

A term number is a recommended position inside a six-month enrollment term, not a promise that a course begins on a universal date or lasts a universal number of weeks. Students usually work through courses in sequence, but acceleration, transfer credit, course availability, field requirements, and mentor decisions can change the order. The live Degree Plan is the student's blueprint.

Use the grid in three passes. First, mark courses already satisfied or transferred. Second, identify courses with external constraints—field placement, classroom access, certification exam, employer project, or a capstone dependency. Third, classify the live assessment instrument from the Course of Study so PA drafting and OA preparation can run on different tracks.

Do not convert the table into a calendar by dividing CUs across weeks. Competency units express academic value, not report length, task count, or exam date. A small-CU course can carry a complex applied artifact, and a larger course can be assessment-heavy in a completely different way.

Run the program as one connected system

Maintain a program ledger with the course code, current instrument, dependencies, target date, evidence needed, and next action. For written work, add the approved problem, audience or learner population, key definitions, data choices, and evaluator feedback. For exam work, add the preassessment result by competency, practice dates, and go-or-wait decision.

The ledger prevents local success from creating downstream rework. A design decision, dataset, learner analysis, security assumption, or framework interpretation introduced early should not silently change in a later capstone. When a course genuinely requires a new premise, record the reason and update every dependent artifact.

Flat-term tuition makes idle time expensive, but acceleration has to stay evidence-led. Keep one PA and one OA-prep lane moving where the Degree Plan permits, front-load externally constrained work, submit only after every rubric aspect is visible, and schedule a proctored exam only after practice shows stable readiness.

Program → class → verified assessment

This page is the program layer. Each linked code opens the class layer with course-specific writing, evidence, data, and competency guidance. The final layer is a public assessment manual, but WGU keeps most task identities inside the authenticated Course of Study. Publishing a generic Task 1 would create the same phantom-assignment problem the network's verification rules were designed to stop.

The WGU rule is therefore strict: at most one PA manual per course, and only when a current WGU-controlled public source verifies its identity and requirements. Objective assessments remain preparation-only. Students sit every proctored exam personally, and tutors never request or use portal credentials.

Quality gates before a course is marked complete

Use three different completion tests because a WGU course can ask for fundamentally different proof. A written PA is ready when every current rubric aspect has an evaluator-visible answer, the evidence supports the nearby claim, the required template and file type are correct, and a final read can trace the conclusion back to facts or analysis. An OA is ready when practice evidence is stable across the tested competencies, not merely when one familiar question set has been memorized. Applied work is ready only after the real activity, approval, hours, documentation, and professional obligations are complete.

Record returned work as structured evidence rather than as a general setback. Put each evaluator comment beside the affected rubric aspect, diagnose whether the gap is coverage, explanation, evidence, calculation, format, or source use, and revise the smallest complete unit that resolves it. Then run a regression pass across dependent sections and files. A changed assumption in a design document, for example, may alter code, tests, diagrams, and the capstone narrative even if the evaluator named only one location.

At the program level, “done” therefore means more than a checked course tile. Keep a compact completion record: official code and title, live assessment instrument, version or date of directions, submission outcome, retained feedback, and any definition, dataset, learner population, framework, or assumption that later courses may reuse. This creates continuity without treating an old task as the specification for a new one. The current Course of Study always wins when a course changes.

Close each term by reconciling that record with the Degree Plan and the next registered course.

Applied and professional responsibility

The student remains responsible for the real case facts, code, calculations, research choices, team participation, professional decisions, originality, and final submission. Tutoring supports learning, planning, and revision without impersonating the student.

De-identify student, minor, employee, customer, and organizational information before sharing any artifact. A tutor can help trace claims to evidence, test whether a method fits the question, reconcile numbers, improve structure, and prepare for an exam. A tutor cannot create events that did not happen or complete an authenticated assessment.

Questions about this program

Keep going