I build and document cloud infrastructure on AWS.

For smart-city operators responsible for sensitive citizen, personnel and IoT data, this framework replaces public exposure and ad-hoc controls with an auditable security model. In the project assessment, two Critical risks scored 9 were reduced to Low, while a three-node database design improved planned service resilience.

For clinical screening and financial analytics teams, this pipeline turns imbalanced, unstructured datasets into decision-ready risk and customer insights. In benchmark testing, tuning lifted stroke recall from 0% in the misleading high-accuracy baseline to 80%, while segmentation produced four actionable customer groups.

For a growing two-site business, this design removes single points of failure, fragmented addressing and insecure inter-site traffic. It provides capacity for 50% headcount growth without re-addressing, uses gateway redundancy to minimise downtime, and passed 18 end-to-end validation tests.

For a regulated digital-services team, this architecture replaces manually managed, single-instance infrastructure with a secure and scalable AWS operating model. Multi-AZ services and a 2–4 instance Auto Scaling range improve continuity and resource utilisation, while central identity and reusable bootstrapping reduce repeat configuration effort and access risk.

For content and commercial analysts, this dashboard removes the need to download raw files or write one-off queries before answering routine catalogue questions. It converted 8,807 records into six decision-ready visuals in roughly two hours, creating a faster and repeatable path from source data to insight.

For retail banking service teams and customers, this prototype shifts repetitive greeting and balance enquiries toward secure self-service. It demonstrates automated handling across two core intents, captures account context, and introduces date-of-birth verification before sensitive information is returned.

For teams managing sensitive or regulated records, this control gives the organisation direct authority over who can decrypt its database rather than relying only on provider-owned keys. The test proved unauthorised reads were blocked and limited approved access to five explicit cryptographic actions, reducing data-exposure and audit risk.

For a hospitality platform that needs repeatable environments, this replaces console clicking with a version-controlled CloudFormation template that a GitHub Actions workflow deploys on demand. Every change is reviewable in a pull request and re-running the pipeline updates the stack in place, so environments stop drifting and rollbacks are a git revert away.