C9 Insight Engine: Pro-Level AI Coaching

Architecture Diagram
Flow Diagram
Playstyle Summary
Game Stats Overall
Achievements
Progress Chart
Your Coach Tip
Highlight Story
Season Headline

💡 Inspiration

The gap between professional esports analytics and the average player’s experience is massive. While teams like Cloud9 have access to dedicated analysts and high-fidelity GRID telemetry, the everyday player is stuck with basic KDA stats that don't tell the full story.

We were inspired to build C9 Insight Engine to democratize pro-level data. We wanted to move away from "Hard work and practice for old ages" and move toward intelligent, data-driven growth. We envisioned a "Spotify Wrapped" style experience that doesn't just show you what happened, but explains why it happened and how to adapt.

🛠️ How We Built It

We engineered a cloud-native serverless pipeline designed for speed and scalability:

The Backend: Built on AWS Lambda (Node.js 22) and API Gateway. We focused on creating a lean data-normalization layer to map complex game telemetry into digestible JSON structures.
The Brain: We utilized Amazon Bedrock (Claude 3 Haiku) to perform sentiment and pattern analysis on match data. By feeding raw stats into Bedrock, we generate natural language "narrative fragments" and coaching tips.
The Frontend: A high-performance React 19 application styled with Framer Motion to provide an immersive, animated UI that makes data feel like a story.
JetBrains Integration: We utilized JetBrains Junie as our primary AI coding agent. Junie was instrumental in refactoring our API logic and handling the complex state-management required to bridge the gap between Riot's APIs and the GRID data structures.

🚀 Challenges We Overcame

Data High-Fidelity: Processing raw match frames is computationally expensive. We solved this by implementing an S3-based short-retention lifecycle, ensuring we only process what is necessary for the AI to generate insights, keeping the system "Green" and cost-effective.
Contextual Logic: It’s easy to calculate a win rate, but hard to define an "Archetype." We had to build logic that looks at and relative to match duration:
Time Constraints: Building a full-stack AI tool in a hackathon setting is a race. Leveraging AWS Amplify for CI/CD allowed us to deploy updates instantly while Junie handled the boilerplate code.

🧠 What We Learned

We learned that Data is the new Coach. Moving from standard APIs to a focus on professional telemetry taught us that the most valuable insights aren't in the final score, but in the "micro-trends" of a match. We also discovered how powerful a Serverless + AI Agent workflow can be for rapid prototyping under pressure.

🗺️ What's Next for C9 Insight Engine

Deep GRID Integration: Moving fully into the GRID GraphQL ecosystem to provide frame-by-frame mechanical analysis.
Live Event Overlays: Bringing these AI narratives to live broadcasts so fans can see "Archetype shifts" in real-time during C9 matches.
Predictive Drafting: Using Bedrock to suggest champion picks based on the historical "narrative" of an opponent's playstyle.