AI Race Strategist: The Digital Twin for F1 Engineering

AI Race Strategist is an autonomous Digital Twin for F1. It bridges IoT telemetry with Atlassian Forge to predict failures, audit FIA compliance via Rovo, and auto-fix hardware via Bitbucket.

Comment

💡 Inspiration: When Milliseconds Matter

Formula 1 teams make split-second decisions based on millions of data points. But in software engineering, when a critical infrastructure failure happens, the response is often slow: manual log analysis, ticket creation, and hours of debugging.

What if industrial engineering had the digital reflexes of an F1 Pit Crew?

Inspired by the partnership between Williams Racing and Atlassian, we built AI Race Strategist — an autonomous agent that doesn’t just monitor failures, but predicts, documents, and fixes them in real time.

🏎️ What It Does

The AI Race Strategist is an autonomous “Ghost Monitor” that lives inside Jira.
It bridges physical hardware (IoT telemetry) and software operations, acting before engineers even realize something is wrong.

While we use an F1 car as our primary use case, this architecture is a blueprint for any Specialized Industry—from aerospace manufacturing to renewable energy grids—where physical sensor data must trigger immediate, software-governed actions.

🔍 Detects Anomalies

Autonomous Incident Triggered

The system ingests simulated IoT telemetry and continuously computes an Anomaly Score \( S(t) \):

$$ S(t) = \frac{|f(t) - f_{baseline}|}{\sigma_{max}} $$

When the threshold is crossed:

$$ S(t) > 0.65 $$

the autonomous incident protocol is triggered.

🧠 Analyzes Root Cause

A trained TensorFlow.js model produces human-readable diagnoses such as:

“Strut thickness insufficient for current aerodynamic load.”

🛠️ Autonomous Repair

Bitbucket Pull Request

The agent automatically opens a Bitbucket Pull Request and computes corrective parameters:

$$ T_{req} = T_{current} + \Delta_{safe} $$

🔐 Enforces Compliance

All fixes are validated against a regulatory rulebook:

$$ \text{If } T_{req} > 8.0\,\text{mm} \Rightarrow \text{Trigger Compliance Warning} $$

📝 Automated Documentation

Post Mortem Confluence

A complete Confluence Post-Mortem Report is generated automatically.

⚙️ How We Built It

We built a full-stack Atlassian Forge app spanning the entire Atlassian ecosystem.

🎨 Frontend

  • Custom React dashboard
  • Embedded Three.js Digital Twin
  • Real-time 3D stress visualization inside Jira

🤖 AI Core

  • Local TensorFlow.js model
  • Predicts failure probability from live telemetry
  • Optimized for low-latency inference

🧩 The Brain (Rovo)

ROVO

  • Built as a Specialized Domain Expert using Atlassian Rovo
  • Understands physics + compliance, not just Jira data
  • Answers questions like “What is the status of the fleet?”

🔁 Orchestration

  • Forge Triggers listen for Jira labels like power-unit-vibrations
  • These labels activate the autonomous remediation pipeline

🚧 Challenges We Ran Into

❌ The Problem

Forge OAuth lacks granular Bitbucket scopes for autonomous file editing.

😤 The Struggle

  • 401 Unauthorized errors
  • OAuth scope mismatches
  • CORS limitations

✅ The Solution

A custom OAuth2 bridge using:

  • bitbucket-oauth-v2
  • Explicit repository:write scopes in manifest.yml

The first AI-generated Pull Request was the highlight of the hackathon.

🏅 Accomplishments We’re Proud Of

  • 👻 Sub-second Ghost Monitor reactions
  • 🧬 3D Digital Twin embedded in Jira
  • 🗣️ Rovo Agent that feels like a teammate

🧠 What We Learned

⚠️ AI Needs Guardrails

Confluence-based Rulebook

In Specialized Industries (Medical, Automotive, Aerospace), safety is non-negotiable.
We implemented a Regulatory RAG, linking Rovo to a Confluence-based Rulebook, ensuring the AI can never bypass compliance constraints.

☁️ The Power of Forge

Forge enabled us to host frontend, backend, and storage entirely inside Atlassian Cloud with minimal operational overhead.

📈 Impact

AI Race Strategist transforms Atlassian from a project tracker into a Mission Control for physical assets.

  • Time-to-Resolution: Reduced from hours to seconds
  • Compliance Accuracy: 100% verification via Rovo–Confluence bridge

🚀 What’s Next

  • 📡 Live MQTT ingestion from ESP32 sensors
  • 🧠 Fleet-wide learning and auto-remediation
  • 🧭 Atlassian Compass integration

AI Race Strategist — Bringing Formula 1 reflexes to DevOps infrastructure.

Built With

Share this project:

Updates