Back to home

E-Learning / EdTech

E-Learning Platform at Scale

Built and operated an e-learning platform serving thousands of daily users, handling full infrastructure on GCP.

Full-Stack Developer & Infrastructure LeadIndependent Contract, Paris2024Team of 10
🎓

Overview

Built an e-learning platform from the ground up that grew to serve 3,000-7,000 daily active users. Handled everything from the application code to the cloud infrastructure, while managing two parallel remote development teams.

The Problem

The client needed an e-learning platform that could scale to thousands of concurrent users without breaking the bank on infrastructure. Previous attempts with off-the-shelf LMS solutions hit performance walls at ~500 concurrent users and couldn't support the custom learning flows the client required, including adaptive quizzes, real-time progress tracking, and a content authoring system for non-technical instructors.

Approach

Architecture for scale on a budget

Designed a Next.js application deployed on GCP with a PostgreSQL database. Used server-side rendering for initial page loads (critical for SEO on course pages) and client-side data fetching for interactive elements like quizzes and progress dashboards. Static content (videos, PDFs) served through Cloudflare R2 to keep bandwidth costs low.

Lead generation and automation suite

Beyond the platform itself, developed a lead generation system with automated scraping, offer publication, and notification pipelines that directly boosted the client's enrollment numbers. This was a separate service layer that fed into the main platform's user onboarding flow.

Infrastructure as code on GCP

Set up the entire infrastructure on Google Cloud Platform: compute instances, managed PostgreSQL, load balancing, CDN configuration, and monitoring. Automated deployments with CI/CD pipelines, so the team could ship multiple times per day without manual intervention.

Challenges

Handling traffic spikes during course launches

Course launches would spike concurrent users 5-10x above baseline within minutes. Implemented auto-scaling policies on GCP compute instances and added a caching layer for course content that could serve stale-while-revalidate during high load. The platform handled every launch without downtime.

Coordinating two remote teams across time zones

Managed two parallel teams of 5 developers each, working on different parts of the platform. Established async communication patterns, clear API contracts between teams, and weekly architecture sync sessions to prevent integration issues.

Results

The platform scaled from zero to thousands of daily users within months, with high availability and low infrastructure costs.

7k

Peak daily active users

10

Developers managed across 2 teams

99.9%

Uptime during course launch spikes

Tech Stack

Next.jsSSR for SEO-critical course pages, CSR for interactive learning flows
Node.jsShared language across frontend and backend, fast iteration
PostgreSQLRelational data model for users, courses, progress tracking, and analytics
GCPFull cloud infrastructure: compute, managed DB, load balancing, monitoring
Cloudflare R2Cost-effective object storage for videos and static course content
VitestTest suite for critical paths: enrollment, quiz scoring, progress tracking

Leadership

  • Managed two parallel remote teams of 5 developers each across different projects.
  • Defined API contracts and integration patterns between teams to prevent merge conflicts.
  • Set up CI/CD pipelines enabling multiple daily deployments with zero-downtime releases.