Phase 6: Architecture defense¶
Present an 8–12 minute, ten-slide defense of your design to a simulated review board. Support decisions with evidence from your portfolio. Distinguish demonstrated, simulated, proposed and untested claims throughout; confidence is not evidence.
| Slide | Required argument |
|---|---|
| 1. Purpose | Stakeholders, problem, intended use and success criteria |
| 2. Model | ERD, grain, keys, relationships and one justified modeling decision |
| 3. Workloads | OLTP/OLAP placement and tradeoffs; explain measurements needed before claiming scalability. Analytics placement depends on workload and isolation requirements. |
| 4. Lifecycle | Collection through retirement, derived copies, validation and refresh/deletion responsibilities |
| 5. Governance | Owners, handling classification, confidentiality impact, regulatory applicability, permitted uses and retention; distinguish policy from enforcement |
| 6. Access | Tested role boundaries, masking limits, and proposed transport/storage protections |
| 7. Monitoring | Actual observations, simulated alerts, gaps in server evidence and response ownership |
| 8. AI and risk | Bias mitigation tradeoff, transparency, appeal, accountability and one harm scenario; state facts needed before making legal claims |
| 9. Infrastructure | Compare cloud, on-premises and hybrid, including residency, sovereignty and cost; justify choice and acknowledge untested availability/recovery claims |
| 10. Decision | Recommend proceed, conditional proceed or defer; name conditions, next checks and unresolved risks |
Submit the presentation and portfolio privately. Text equivalents for visual diagrams and readable tables are welcome. Use accessible text evidence; screenshots are optional. Include the peer review response and individual reflection.
Board questions¶
Be prepared to explain one alternative you rejected, one control not yet tested, one stakeholder who bears risk, and one result that would change your recommendation. Connect each design decision to its requirement, implementation status, evidence and limitations. Do not infer legal consequences or production readiness from the synthetic exercise alone.
Assessment¶
Assess architectural reasoning, evidence quality, cross-phase consistency, tradeoffs, AI lifecycle accountability and clear communication. Use the public descriptors; official grade weights remain in the LMS. A justified decision to defer deployment can be strong work.
Author: Isaac K. Nti.