AWS Cloud Foundation Uplift (Transport Operator Melbourne)
Role: Principal Cloud Architect
Focus: AWS platform uplift (security, identity, network, DevOps/IaC, cost)
Deliverables: Target state + roadmap + governance decision pack
Success metric: Enterprise Risk Reduction, Operational maturity and Cost Optimisation
Context
A major metropolitan transport operator was running several business-critical digital applications on AWS (e.g., operational scheduling, workforce and customer-facing systems). A third-party review identified material security and governance gaps, inconsistent operational practices, and limited repeatability across environments.
With constrained funding and high operational impact, the organisation needed a pragmatic uplift: reduce the highest risks quickly, establish clear platform standards, and create a realistic roadmap to improve AWS maturity over time.
My role
As a Principal Cloud Architect, I was accountable for:
- Assessing the current AWS footprint and risk posture
- Defining a target-state cloud platform architecture (security, identity, network, observability, DevOps/IaC, cost)
- Creating a phased uplift roadmap and aligning stakeholders through architecture governance
This case study focuses on the AWS cloud platform uplift component.
Approach and delivery
1) Rapid current-state assessment (Weeks 1–4)
- Ran focused workshops with the external assessor, cloud vendor, DevOps, security, and network teams
- Mapped findings to the existing platform and application landscape, separating urgent remediation from strategic uplift
- Produced a prioritised backlog (severity tiers) covering security exposures, operational fragility, and delivery gaps
2) Target-state architecture (security + delivery + cost lenses)
Anchored to AWS Well-Architected principles, I defined a target state that balanced ambition with operational reality, covering:
- Security & governance: centralised audit logging and posture visibility; hardened access paths; encryption baselines and secrets handling
- Identity & access: role-based access, least privilege, clearer joiner/mover/leaver controls and controlled elevation
- Network segmentation: environment separation and tighter traffic boundaries to reduce blast radius
- Observability: standardised telemetry and alerting foundations to support faster incident triage
- DevOps repeatability: consistent build/deploy patterns to reduce drift and manual change risk
- Cost guardrails: tagging, budget visibility, and operational levers for non-production environments
3) Phased roadmap + architecture approval
- Produced 2–3 staged roadmap options, sequencing uplift by risk reduction value, delivery effort, and operational constraints
- Built a decision pack articulating trade-offs, benefits, dependencies, and sequencing
- Presented and iterated with senior stakeholders (technology and business leadership) and obtained architecture governance approval
Outcomes
- Established a clear, approved target state and a phased uplift roadmap aligned to risk priority and funding reality
- Enabled immediate remediation of the highest exposures while setting the platform up for sustained maturity uplift
- Improved stakeholder alignment by translating audit findings into a structured uplift program with clear governance
Key learnings
- Start with a risk-to-control map, not a tool list
- Adoption beats perfection
- Governance must accelerate delivery, not block it
- Security posture is mostly identity + visibility
- Network segmentation is your blast-radius control
- Repeatability is a security control
- Observability is a platform feature
- Cost controls need to be designed-in
- Phased roadmaps outperform big-bang programs
- The hardest part is stakeholder alignment
Final thought
Platform uplift succeeds when it’s engineered for adoption. By translating audit findings into a prioritised backlog, aligning stakeholders through decision packs, and rolling out repeatable patterns in phases, we reduced risk quickly while building momentum toward a sustainable target state. The highest-leverage changes came from identity, visibility, and standardised delivery—rather than adding more tools.
Note: details have been generalised for confidentiality.