English Learning Platform

How Unmaze moved to the cloud without breaking under growth

Unmaze, a fast-growing English tutoring platform, had a fully built application that wasn't in the cloud, and it was showing performance and stability issues under traffic. We migrated it to Google Cloud Platform with serverless infrastructure — eliminating the instability and cutting costs to zero during idle periods.

Tell us a bit about your setup and we'll follow up to schedule a 15-minute call.

Senior-led engagement, from discovery to handover.

Certified by:

At a glance

Client: Unmaze
Industry: EdTech — English language learning
Services: Migration
Stack: Google Cloud Run, Firebase Hosting, Docker

Business Challenge

A rapidly growing English tutoring startup faced critical technical challenges that threatened its expansion:

Fully developed application, but not in the Cloud

Performance and stability issues at high traffic

Our Solution

We carried out a complete infrastructure modernization on Google Cloud Platform:

Cloud Migration Strategy

Scalable backend architecture with automatic scaling via Google Cloud Run
Frontend deployment on Firebase Hosting for optimal performance
Complete transition to serverless architecture with custom Dockerfiles for development and production

Database Optimization

Continuous performance monitoring in real-time
Identification and resolution of database bottlenecks

Cost Efficiency and Reliability

Strategic database resizing to optimize costs
Geo-replicated automated backup system

Results

Moving to a serverless architecture on Google Cloud resolved the platform's stability issues and changed its cost profile.

Zerocost during inactive periods, with minimal frontend spend

Eliminatedinstability from high database connection volume

Day onefull cloud benefits leveraged from the first day in production

Frequently asked questions

Why was Unmaze not in the cloud before this migration?

The application was fully built, but had never been deployed to cloud infrastructure — it was running in a way that couldn't reliably handle growing traffic.

How does the platform have zero cost during inactive periods?

We deployed the backend on Google Cloud Run, which scales to zero automatically when there's no traffic, so Unmaze doesn't pay for idle capacity.

What caused the database instability, and how was it fixed?

High connection volume was overwhelming the database. We identified the bottleneck, resized the database strategically, and set up continuous performance monitoring to catch issues early.