Building LIVE Solution: Real-Time Information for Millions of Travelers

Every day, millions of passengers pass through French train stations. They need one thing above all: accurate, real-time information. That's what LIVE Solution delivers, and I had the privilege of helping build it at SNCF Connect.

The Scale of the Challenge

Imagine building a system that needs to: - Process thousands of real-time updates per minute - Display information across hundreds of screens in major stations - Handle peak loads during rush hour - Never go down (seriously, never)

That was our mandate. And we built it on AWS, using a serverless architecture that could scale automatically.

The Architecture

We chose AWS Lambda for compute—functions that spin up on demand and scale automatically. TypeScript was our language of choice, bringing type safety to our serverless functions. The frontend was built with React, creating responsive interfaces for multiple types of displays.

Working in a SAFe Environment

This wasn't a small team project. We had 42 developers organized into 6 squads. I worked within the SAFe framework, participating in PI planning, sprint reviews, and continuous delivery. Coordination was key—one squad's work often depended on another's.

The Testing Pyramid

I'm passionate about testing, and LIVE Solution became a showcase for comprehensive test coverage: - Unit tests for every function - Integration tests for API endpoints - End-to-end tests using RobotFramework with Python - Performance tests to ensure we could handle peak loads

We lived and breathed TDD/BDD. Write the test first, watch it fail, make it pass, refactor. This discipline saved us countless hours of debugging.

Quality Gates Everywhere

Every commit went through SonarQube analysis. Code coverage below 80%? The pipeline failed. Cyclomatic complexity too high? Refactor required. We used OctoPerf for load testing, simulating thousands of concurrent users.

Monitoring and Observability

In production, we relied on CloudWatch for AWS metrics and DataDog for application-level monitoring. We built dashboards that showed everything from Lambda execution times to error rates to user engagement metrics. When something went wrong, we knew about it before users did.

The AWS Ecosystem

We leveraged AWS deeply: - Lambda for compute - DynamoDB for lightning-fast data access - S3 for static assets and data storage - Step Functions for orchestrating complex workflows - CloudWatch for monitoring and alerting

The Human Element

Technology is only half the story. The other half is people. We worked closely with station staff, understanding their needs, iterating on designs based on feedback. We shadowed them during peak hours, watching how they used our system.

The Impact

Today, LIVE Solution powers passenger information across France's major train stations. Millions of travelers rely on it daily. The system handles peak loads effortlessly, and our monitoring shows 99.9%+ uptime.

What I Learned

Building at scale requires more than technical chops. It requires: - Robust architecture that can handle failure - Comprehensive testing at every level - Monitoring that gives you confidence - Collaboration across large, distributed teams - Never compromising on quality, even under pressure

The Proudest Moment

Standing in Gare de Lyon during rush hour, watching thousands of passengers glance at screens powered by code I helped write, seeing them make confident decisions about their journeys—that's why we build systems.