Full Stack / DevOps Developer

Arvison

Unreal streaming on AWS · NestJS · matchmaker and WebRTC signaling

Designed and built a cloud streaming platform on AWS that enables users to access GPU-intensive Unreal Engine applications from any device. Built an optimized NestJS server for Unreal Engine connectivity, a matchmaking system for dynamic instance allocation, and scalable infrastructure using WebSocket.IO and WebRTC for low-latency streaming.

AWS
Nest.js
Node.js
Socket.io

AWS: Scalable cloud infrastructure
WebRTC: Low-latency streaming
EC2: On-demand Unreal Engine instances
Matchmaker: Dynamic session assignment

The story

01.1

Problem

GPU-intensive Unreal Engine applications demand powerful hardware, keeping most users from running them on their own devices.

01.2

Solution

I architected a streaming platform on AWS: a matchmaker dynamically assigns EC2 instances running Unreal Engine, and video is delivered over WebRTC with WebSocket.IO signaling.

01.3

Result

Applications became accessible from any device with a browser, with automatic shutdown of idle instances to cut operating costs.

What I built

02.1Devops
Design and Development of Infrastructure on AWS
Contribution to the design and development of an AWS infrastructure to generate instances of Unreal Engine applications accessible through a matchmaker, enhancing the system's accessibility and scalability.
02.2Devops
EC2 Image Generation
I generated custom images on AWS EC2 for the matchmaker and Unreal Engine instances, optimizing deployment and operational efficiency.
02.3Full Stack Developer
Development of Signaling Server
I programmed functionalities in Node.js for the automatic control of instances without connected clients, reducing operational costs.

The code for this project belongs to the client and cannot be published. What I can show is on this page: the problem, the solution I built, my specific role and screenshots of the product. If what you want is to read my code, that's exactly why I built Ghostline: a complete, playable game with both the C# client repo and the TypeScript API repo open —the same technical judgment I applied here, this time auditable commit by commit.

Gallery

**AWS Architecture**Infrastructure diagram: the user enters through the load balancer, the matchmaker assigns an EC2 instance running Unreal Engine, and video is delivered over low-latency WebRTC.

**Session Lifecycle**From request to automatic shutdown: matchmaking queue, streaming, idle detection, and session end with instance release.

Next project

Insurtech Ecosystem

Insurance platform · Web, mobile and CRM

Back to the cases Insurtech Ecosystem →