R6 Stage D · 160 h (64 T / 96 P) Advanced
AI Solutions Architect
Designs AI systems that satisfy functional, economic, security and regulatory constraints.
Entry requirement: Gates 1–3 at ≥80% plus demonstrated communication competence.
Tech stack
AWS BedrockAzure AIGoogle Vertex AITerraformAPI gatewaysLiteLLMIAM/KMS
Modules
M1 · Architecture method
40 h TOPICS
- Requirements and constraint capture
- Quality attributes
- Trade-off analysis
- Architecture decision records
- Reference architectures
- Build versus buy
- Managed versus self-hosted
- Vendor lock-in and exit strategy
Lab: Produce an architecture decision record with a defended trade-off analysis
Course material for this module is in production.
M2 · Designing for scale and resilience
40 h TOPICS
- Capacity planning
- Throughput modelling
- Quota and rate-limit architecture
- Multi-region and failover
- Graceful degradation
- Fallback model chains
- Queueing and asynchronous patterns
- Disaster recovery objectives
Lab: Highly available design operating under provider quota limits, with failure-mode analysis
Course material for this module is in production.
M3 · AI economics
40 h TOPICS
- Cost drivers and modelling
- Token and inference economics
- Caching tiers
- Right-sizing
- Reserved versus on-demand
- Storage and egress
- Total cost of ownership
- Chargeback and FinOps practice
- Cost guardrails
Lab: Build a defensible cost model and TCO comparison for a stated workload volume
Course material for this module is in production.
M4 · Security, compliance and governance architecture
40 h TOPICS
- Tenancy isolation
- Network and identity design
- Encryption and key management
- Data residency and sovereignty
- DPDP Act obligations
- Auditability
- Model risk governance
- Third-party assessment
- Security review process
Lab: Complete a security and governance design review against a control checklist
Course material for this module is in production.
Track project
Enterprise AI reference architecture with a working reference implementation: architecture decision records, scaling and resilience plan, quantified cost model, security and governance review, and an executive presentation defending the design.
Job-ready exit standard
Designs and defends an AI system end to end; cloud architect or AI certification recommended.