Skip to main content

Workload and progression

Plan 510–725 hours of focused work, including repairs. These are author estimates requiring learner timing pilots. The legacy Weeks 76–83 sequence is an ordering aid, not an eight-week completion promise. At 20 hours/week this is approximately 26–37 weeks; at 30 hours/week, 17–25. Peer scheduling and delayed transfer can extend calendar time.

Allocation without double counting​

  • Module 01: 80–115 hours, comprising original work 60–85 and capacity package 20–30.
  • Module 02: 85–120 hours, comprising original work 63–88 and migration package 22–32.
  • Module 03: 90–125 hours, comprising original work 68–93 and saga package 22–32.
  • Module 04: 85–120 hours, comprising original work 65–90 and burn-window package 20–30.
  • Module 05: 80–115 hours, comprising original work 62–89 and leadership package 18–26.
  • Required product/platform/AI/sustainability strand: 40–55 hours.
  • Common statistical-learning bridge: 20–30 hours (concepts 5–7, experiment 7–11, error/regularization 4–6, assessment 4–6).
  • Semester integration, defense and repair: 30–45 hours.

Modules total 420–595 hours. The integration allocation includes project assembly 12–18, checkpoint and cumulative retrieval 5–8, exam 2.5 plus defense 2.5, and repair 8–14. Module implementation, readings, quizzes and ordinary practice are already included. Strand work reuses module artifacts and charges only its added investigation; project assembly must not repeat entire implementations.

Sequence and preserved depth​

Complete capacity before migration, migration before saga/replay, and actual reliability instrumentation before the leadership decision. Develop the project throughout. Retain the 45-minute unfamiliar design exercise, two case-study deep dives, actual extraction/storage migration, durable outbox/idempotent consumption with crash cuts, three-service saga, bounded poison replay, actual load/overload and incident experiments, two pre-reads and peer delegation. Small Python contract models are bridges to these exercises, not substitutes.

Use one synthetic-data local service across modules where practical. Record offered, admitted and completed work; distinguish predicted capacity from measured throughput. Show the original durable store and broker failure behavior. A finite model does not prove deployment fencing, durable compensation, full alert routing or production-scale operation.

Conjunctive progression gate​

Score 0 missing/incorrect, 1 prompted/incomplete, 2 independently supported with limits, 3 fresh transfer. Every SD1–SD6, MC1–MC6, EV1–EV6, RE1–RE6 and TL1–TL6 category must reach at least 2. Every P1–P5 strand category and SL1–SL4 bridge gate must also reach at least 2. A high average cannot compensate for a missing gate.

Additionally satisfy the existing checkpoint minimum 80%, exam minimum 70% overall and 60% per major section, all project deliverables and integrity requirements. Conditional pass means repairs are pending; it does not authorize progression. Preserve the module-specific delayed transfers at least 48 hours later. Actual pre-read and delegation evidence are mandatory; rehearsals do not invent participation.

Evidence packet and handoff​

Index each gate to raw evidence, implementation revision, environment, independent check, limitation and repair. Link the design, estimates, migration cutover/abort evidence, saga outcomes/replay runbook, SLI/alert/load/incident traces and reviewed strategy. Retain P1–P5 evidence beside these artifacts. Sanitize identities and secrets before public portfolio use.

Carry the highest unresolved operational risk into Semester 09: what must be automated, how credentials/configuration are controlled, which deployment or restore test could falsify readiness, and who owns the result. Do not claim a production deployment from a local model. Author harness checks validate teaching artifacts; learner service runs, interpersonal work and timing pilots are separate.

Assistance and verification evidence​

Complete the linked review exercise as required subchecks of P4, using 2–3 hours of the existing 9–13-hour AI-risk allocation. Reuse the existing packet; no extra semester hours or gate IDs are added. All original real-system and independent-review requirements remain.

Common statistical-learning requirement​

Complete the statistical-learning bridge before P4. SL1–SL4 each require level 2, alongside all existing gates. Its 20–30 hours are additional to P4 and included in the revised 510–725-hour total. Reuse its evidence in integration; no additional timed exam section is introduced.

Independent practice route and evidence reuse​

The following entry checks and changed-case practice make the existing module work inspectable. Use them within the current original-practice, lab and retention allowances; no full-book reading or new parallel project is added. An existing artifact can satisfy both an original exercise and this route when it demonstrates the same claim. Keep one evidence index linking each required gate to the exact artifact, actual run, assistance and reviewer status. Do not waive a topic simply because a different lab passes.

Keep authored material, author execution, learner attempts, independent review and delayed transfer as distinct statuses. Retain existing construction, security, HCI, AI-use, world-class and capstone strands and all per-gate minima. Real-system and human requirements remain pending until observed; planned observations are not completed evidence. Record actual study time before revising estimates.