Selected work · Status made explicit
This portfolio connects professional FinOps and TBM practice with developing work in AI economics, policy automation, and lightweight software. Each entry separates current maturity from intended value.
Portfolio standard
The common thread is a governed system: business objectives shape policy; policy guides architecture, access, and accountability; usage and cost evidence inform optimization; and delivery lessons improve the governing standards.
Intended outcomes are not presented as realized results when supporting evidence is not yet available for public review.
Portfolio
Maturity ranges from established professional practice to early concepts. Status and evidence limitations are part of each project description.
Problem and approach. Cloud visibility alone does not create ownership or a repeatable path from financial signals to action. Apply FinOps and TBM methods to governance, allocation, forecasting, optimization, persona reporting, and organizational adoption.
Evidence boundary. This reflects capability-level professional practice. Client identities, proprietary materials, and unpublished outcomes are excluded.
Problem and approach. Stakeholders often struggle to translate decisions into durable cloud reports. A conditional interview maps persona, decision need, and priorities into a best-practice report specification and automation-ready definition.
Evidence boundary. Designed to reduce requirements churn. Reliable production report creation across Cloudability environments has not yet been demonstrated.
Problem and approach. Billing exports omit object-level attributes needed to estimate one-time storage-transition charges precisely. Combine S3 Inventory metadata, pricing, and storage rules to evaluate small objects, transition eligibility, and minimum-duration conditions.
Evidence boundary. Calculation logic remains under validation and is not represented as production-accurate until reconciled against authorized inventory and billing evidence.
Problem and approach. AI access is commonly governed at the account level even though value, cost, and risk depend on role and task. Map job function, seniority, use case, and risk to permitted model candidates and an explainable recommended fit.
Evidence boundary. This is a governance framework, not a deployed CMDB integration or autonomous policy engine.
Problem and approach. Standards can become static while teams discover new risks and better delivery patterns. Capture release evidence, convert it into a policy proposal, route review and approval, and record the resulting change or exception.
Evidence boundary. Templates, structured data, and agent-assisted workflows are being explored; this is not a production autonomous governance agent.
Problem and approach. Maintaining business mappings is difficult, while activating untested changes can alter allocation and reporting. Compare current mapping extracts with candidate rules supplied by CSV, with a future CMDB API path where supported.
Evidence boundary. A supported dry-run endpoint has not been confirmed. Feasibility depends on available exports, APIs, and rule semantics.
Problem and approach. Optimization findings remain optional when detached from policy, evidence, ownership, and exceptions. Define service-specific review policies as machine-readable controls, then use read-only agents to evaluate and route findings.
Evidence boundary. Thresholds, data quality, control ownership, and integrations remain unvalidated; no autonomous write authority is proposed.
Problem and approach. General event tools do not closely fit one small golf group’s trip and scoring needs. Build a deliberately narrow mobile application while applying secure delivery, lifecycle governance, and cloud cost controls.
Evidence boundary. This invitation-only MVP is an engineering and governance laboratory, not a public product or primary career credential.
View project detailsConfidentiality and evidence