One intelligent entry point that turns a fragmented stack of cloud-platform tools into a single, calm command center — for engineers and non-technical teams alike.
The Platform Overview — health, cost, pipelines, and alerts in one glance.
During my internship I helped design the Platform Operations Portal (POP) — a centralized interface that lets teams monitor health, track cost, manage data lifecycle, and provision environments without leaving one screen.
Design one entry point to monitor usage, track costs, manage workflows, and provision environments.
An enterprise portal serving both technical and non-technical users — usability and clarity first.
UX design intern — dashboard, provisioning flows, reviewer system, alerts, and an AI interaction layer.
A quick walkthrough of the working prototype — the dashboard, guided provisioning, and the AI interaction layer in action.
A short walkthrough of the working prototype.
The existing ecosystem relied on multiple fragmented tools — leaving users with limited visibility and a maze of systems to cross.
Troubleshooting was slow, decisions stalled, and almost everything depended on deep technical expertise.
"Users shouldn't need to understand the system to use it effectively."
No single place to see platform activity, health, or spend.
Insights and actions were scattered across disconnected systems.
Troubleshooting and approvals dragged, bottlenecked by technical experts.
I focused on turning sprawling operational complexity into clear, guided experiences across five areas of the portal.
A consolidated overview of platform health, cost, and activity.
Step-by-step environment provisioning and approval flows.
A reviewer-selection system that routes requests to teams, not individuals.
A structured alert & notification system that surfaces what's critical.
A prompt-based search that answers questions and triggers actions.
Each solution removes a layer of friction — consolidating tools, adapting to roles, and guiding users through the hard parts.
One dashboard brings together project inventory, cost & usage insights, and pipeline status — replacing a stack of disconnected tools with a single source of truth.
A "View as" model tailors the interface to engineers, analysts, product owners, and governance users — each sees the information and actions relevant to them, nothing more.
Complex tasks like environment provisioning become a calm, four-step flow — Environment Type → Select Services → Configure Resources → Review & Deploy — with clear progress and sensible defaults.
Instead of chasing individuals, users select reviewers by team or role, send requests to a group, and track approval status clearly — so nothing stalls on one person's inbox.
A prompt-based assistant lets anyone ask in natural language — "Why did costs increase?", "Show failed pipelines" — and get traceable, sourced insights, with the option to trigger actions right from the answer.
Every AI insight is "Traceable" — showing its sources, data lineage, and scope & limitations, so users can act on it with confidence.
The design set out to make platform operations visible, faster, and approachable — for experts and newcomers alike.
Consolidated visibility across platform operations, replacing constant tool-switching.
Guided flows and AI insights shorten troubleshooting and approval cycles.
Simplified onboarding makes the platform usable for non-technical teams.
Designing for enterprise means balancing complexity with clarity.
The biggest lesson: users should never need to understand the system to use it effectively. Every flow had to hide operational complexity while keeping people in control.
The final concept transformed a fragmented experience into a cohesive, scalable, and user-friendly platform — supporting both operational efficiency and long-term adaptability.
Note: Based on work completed during an internship. Specific organizational details have been generalized to maintain confidentiality.