Identified 6 systemic gaps between how cloud consoles are built and how DevOps engineers actually work.
Informing the design direction for Huawei Cloud's global mobile console app — through competitive analysis, user interviews, and concept design.
Context
The Project ↗
Huawei Cloud's mission: build a mobile console app for global market users managing ECS (Elastic Cloud Server) infrastructure. The goal was to give DevOps engineers anywhere access, real-time monitoring, and on-the-go troubleshooting.
My Role
I was on the CRC Toronto team, handling the global market side. My contribution covered the full research pipeline: UXR strategy, user interviews, competitive analysis, persona, journey map, and concept design directions.
Research
We ran a 6-step process: competitor analysis → heuristic evaluation → UXR strategy → user interviews → insight synthesis → design directions. I led the global market track.
Overseas competitors (AWS / Azure / GCP)
Clean UI, CloudShell support, essential features covered. But: no advanced management capabilities, no learning community, limited mobile depth.
Alibaba Cloud
Rich learning resources, super-app model. But: information overload, no user onboarding, navigation is hard to parse.
Key gap — All competitors support basic ops. None support a customized, deep management experience for infrastructure-heavy users.
Alibaba Cloud analysis snippet
Key Insights
From user interviews and a mental model exercise, we identified 6 systemic gaps. Three shaped the concept design most directly.
Insight 02
Mental model mismatch — users think in products, consoles are organized by resource type
SRE and infrastructure teams have different mental models, creating high communication overhead. The console's resource-first structure doesn't map to how DevOps engineers actually manage their work.
"I have to jump between 4 different pages just to see if my product's servers are healthy."
Insight 05
Security is the #1 blocker for mobile adoption
Restarting a server means product downtime. Mobile security policies vary by company. Engineers won't act on mobile without explicit safeguards — the risk is too high.
"I would never restart a server on my phone. What if I tap the wrong thing?"
Insight 06
AI-assisted automation is the expected future
DevOps workflows are only ~50% automated. Engineers already use ChatGPT for config help — they expect the console to meet them there, with a human confirmation loop preserved.
"I would use the app to quickly check or fix something. I just want the key alert information and the way to solve my server issue."
Concept Design
Three screens derived from research — each tracing directly back to a specific insight and user goal.
Screen 01
Alert Dashboard
Critical alerts surface immediately with enough context to act without opening a laptop.
Screen 02
Diagnose & Resolve
Root cause context + AI-suggested fix — without switching to desktop.
Screen 03
Project-based Overview
Servers grouped by product, not resource type — matches how SREs actually think.
Reflections
Derive actionable steps from large data
This project taught me how to find meaningful insights from large volumes of data. I learned to conduct mixed-method research — choosing the right approach based on what the project actually needed, not just what was convenient.
Designing with Ambiguity
The project didn't start with a clear goal. It was only through research and conversations with users that we discovered we were designing for the 4am moment — the stressed DevOps.