[GHJ306] AWS GameDay: Developer Experience Powered by AI (Provided by New Relic) #AWSreInvent

Hello.
I am Takaaki Tanaka from the Manufacturing Business Technology Department.

I participated in AWS GameDay: AI-Powered Developer Experience (provided by New Relic), so I'd like to report on the experience.

On the day, after the 5K Run finished, it was a long 4-hour GameDay from 8:00 - 12:00, but it became an extremely valuable learning experience.

Overview

Join this immersive and collaborative learning exercise to gain practical skills with AWS AI-assisted development tools. As a new software engineer at a fictional startup company, you'll dive into a live AWS environment and be able to leverage AI to multiply your engineering skills tenfold. While working with AI-powered coding assistants and other tools, you'll utilize and build your skills by planning prompts, clarifying specifications, confirming requirements, and completing tasks to reach the top of the leaderboard. Race against time to transform legacy codebases, automate operations, and even ship directly to production. Whether you're an experienced coder or just beginning your cloud journey, this GameDay will be the ultimate test of your development skills. This session will cover integration use cases with New Relic. Walk-ups are welcome.

To summarize briefly:

• Setting is as a new engineer hired at a fictional startup
• Hands-on with an actual AWS environment
• Utilizing AI-powered coding assistants
• Completing tasks such as refactoring existing code and automating operations

It's a hands-on type event where you "learn by doing."
This time, integration use cases with New Relic were included, requiring observability and monitoring perspectives.

Impressions of the Work

Since I can't share the specific content and structure of the GameDay, this will be an impression based on the atmosphere.

Conquering the Environment in Quest Format

A pre-prepared AWS environment was distributed, and we solved multiple "quests" in sequence or in parallel.
Each quest focused on themes such as application and infrastructure improvements, troubleshooting, and feature additions. These weren't just simple tasks that could be completed by following instructions or asking AI—they required thinking on our own.

Understanding the Entire Application is Required

In terms of difficulty, it wasn't just about knowing specific AWS services, but also:

• The overall picture of the application architecture
• Dependencies and deployment flows
• How to read monitoring, logs, and metrics

I felt that without the ability to "see the system holistically," it would be difficult.

In particular, if you don't consider how a change might affect other components or quests, a change that seems correct might not lead to points or could cause other problems.

AI Assistance is Excellent...

While AI assistance was assumed for the GameDay, simply delegating everything to AI wouldn't work. I felt that without proper prompt design that provides the right context and the ability to verify AI suggestions with your own experience and knowledge, it would be difficult to achieve score-worthy results.

Especially when modifying old codebases or making infrastructure changes, rather than directly applying code or procedures provided by AI, we needed to proceed while judging for ourselves whether they "match the application's assumptions" or "pose any security problems."

Results

As a final result, we placed 6th (probably) among over 50 teams (probably) participating.

Personally, I felt I couldn't contribute at all, and achieving this ranking was due to the high level and coordination of my team members.

I was blessed with wonderful members, and it was truly a great experience.

Thank you to everyone who participated with me.

For those interested in practical hands-on experience with AI-assisted development tools and AWS, I highly recommend this GameDay.
I felt it allows you to experience the sensation of expanding your skills using AI, rather than leaving everything to AI.