Helping finance and engineering agree on cloud spend.

Cloud cost clarity for finance, DevOps, and cloud admins acting on the same spend picture.

View Cloudi screens in Figma

3primary user roles

AWS · Azure · GCPmulti-cloud scope

Shippedinternal product

Overview

The same cloud bill meant different things to each team.

I shaped research synthesis, information architecture, dashboards, recommendations, alerting, and handoff for an internal AWS cost product used across finance and engineering contexts.

The problem

Visibility was fragmented before action even began.

Different versions of spendTeams reconciled disconnected reports after cost risk was already visible.
Different mental modelsFinance needed confidence, DevOps specificity, and cloud admins control.
Charts without actionThe first dashboard showed spend but buried what users should do next.
Unsafe recommendationsOptimization advice needed confidence, risk, resource evidence, and ownership.

The solution

One path from cost signal to owned action.

See

Give every role a common spend and trend overview.

Explain

Connect anomalies to technical resources and causes.

Act

Pair recommendations with risk, evidence, and an owner.

Decision 01

Create one shared operating view.

The overview gave finance and engineering the same starting point before offering role-specific depth.

Shared language. Use consistent spend, trend, ownership, and risk definitions.
Progressive depth. Move from overview into team, cluster, and resource detail.
Context travels. Keep the selected period and ownership visible across views.

Cloudi dashboard with monthly cost, account spend, and cost categories — The updated home view brought account spend, period comparison, and cost categories into one shared starting point.

Decision 02

Make every recommendation explain itself.

The first version buried recommendations below charts. I moved prioritized actions into the main path and attached the evidence needed to judge safety.

Cause before action. Explain what produced the cost change.
Confidence stays visible. Show the strength of the evidence.
Risk and owner. Make operational consequences and responsibility explicit.

Cloudi Databricks cost recommendation comparing current and recommended cluster configurations — The recommendation detail paired an estimated cost opportunity with the current and proposed cluster configuration.

Decision 03

Connect the anomaly to the resource.

A cost signal became useful when an engineer could trace it to the cluster or resource that required investigation.

Signal to detail. Preserve the route from overview into technical evidence.
Operational state. Keep runtime and resource health visible.
Action in context. Avoid sending users into a separate reporting tool.

Cloudi Databricks cluster overview with job status, expense, and resource usage — The cluster view connected job state and expense with the operational metrics an engineer could investigate.

Outcome

Teams gained a clearer route from cost risk to action.

The shipped internal product created a shared operating view and made recommendations easier to inspect.

Shared cost context

Finance, DevOps, and cloud admins could work from the same underlying view.

Explainable recommendations

Evidence, confidence, risk, and ownership traveled with the suggested action.

Earlier intervention path

The workflow connected spend change to technical cause while action was still possible.

Evidence boundary: stakeholder interviews, cloud-team shadowing, support feedback, usability testing, and shipment of the internal product support this narrative. No savings, adoption, or business-impact percentage is claimed.

Final product

The shared cloud-cost workflow.

The final product screens are grouped here so the visual walkthrough follows the reasoning and outcome.

Cloudi resource dashboard with expense trend and resource table — The resource overview connected cost movement to the accounts and resources behind it.