SYSTEMS · ESPORTS ANALYTICS

Team Intuition Engine

An AI analyst and coach that turns official GRID match data into advanced metrics, what-if simulations, and live strategic guidance for League of Legends and VALORANT.

↓ SCROLL

The problem

Esports teams review matches from raw stat lines that miss context. A KDA tells you who died, not why a round was lost or what a team should have done differently. Coaches need deeper competitive metrics and in-match guidance, not just box scores. Team Intuition Engine pulls official match data from the GRID Esports API and turns it into Moneyball-style analysis for League of Legends and VALORANT.

The approach

A FastAPI backend ingests match data through GRIDClient, then runs it through title-specific processors that compute advanced metrics beyond basic statistics. The structured output is handed to an LLM layer that writes narrative analysis and coaching recommendations, so numbers become readable strategy. A Next.js dashboard called Junie surfaces the metrics and insights and can simulate alternative outcomes from live game state.

System design

Language / runtimePython 3.10+ backend, Node.js 18+ frontend
Backend frameworkFastAPI
FrontendNext.js, React, TypeScript
Data sourceGRID Esports API (api-op.grid.gg)
AI layerLLM orchestration (DeepSeek / OpenAI)
Core servicesGRIDClient, ValorantStatsProcessor, ValorantAnalyzer
DeployRailway (Procfile, railway.json)

How it works

01

Pull official match data

GRIDClient authenticates with a GRID API key and fetches official match data for League of Legends and VALORANT.

02

Compute advanced metrics

ValorantStatsProcessor and ValorantAnalyzer calculate metrics like ACS, KAST %, and Economy Impact that go beyond basic KDA and Combat Score.

03

Generate narrative insight

The structured stats are passed to the LLM layer, which produces readable analysis and strategic recommendations.

04

Simulate scenarios

What-if modeling replays alternative match outcomes from live game state to test hypothetical decisions.

05

Serve the Junie dashboard

The Next.js dashboard renders metrics, insights, and contextual coaching during active matches.

What it does

Advanced analytics

Calculates ACS, KAST %, Economy Impact, and other metrics that basic stat lines leave out.

Scenario simulation

Models hypothetical what-if outcomes by replaying alternative match states from live game data.

Real-time coaching

The Junie Dashboard delivers contextual strategic recommendations during active matches.

Multi-title support

Works across both League of Legends and VALORANT from a single platform.

Narrative insight from stats

An LLM turns raw statistics into plain-language analysis a coach can act on.

Stack

PythonFastAPINext.jsReactTypeScriptGRID Esports APILLM orchestrationRailway
2025Year
FastAPI + Next.jsStack
SoloRole
LoL + VALORANTTitles