Debugging OpenAI with Claude and Codex: Crafting an AI-Powered Solution

The Challenge That Sparked Arise & Shine Transporters

Wambugu was a seasoned project manager but had never tackled a full-stack logistics platform before. He began with the basics of backend Node.js, frontend React, and integrating GPS telemetry APIs like Protrack 365 for sand and aggregate delivery operations in East Africa.

Builder Notes: Arise & Shine Transporters started as simple order management that tracked where trucks were via live GPS data from Protrack 365. The goal was to replace manual phone calls with real-time visibility into truck locations."

\`\` What It Is

Arise & Shine Transporters is a full-stack logistics platform for sand and aggregate delivery in Kenya, enabling the business owner to manage orders, monitor fleet performance, generate cost reports, and automate financial processes. **** Sand trucks on road with GPS telemetry devices"

\`

The Problem That AI Helped Solve

Before Arise & Shine Transporters, everything was done manually — phone calls, paper records. Wambugu encountered many challenges: knowing where the trucks were at any moment, calculating costs accurately, and tracking driver routes.

Builder Notes: Initial version lacked distance pricing logic; it cost KES 30,000 per delivery regardless of distance. This was a significant oversight that threatened the platform's viability."

\`\` The Solution with AI

To address this issue, Arise & Shine Transporters introduced dynamic, configurable pricing models: KES 32,000 base for first 15 km and +KES 1,000 per additional 5km. This was a game changer, making the platform more accurate and cost-effective.

Builder Notes: The AI-powered distance-aware pricing feature used OpenAI's Claude SDK to calculate delivery costs on-the-fly based on real-time GPS data from Protrack 365."

\`\` Debugging with AI Collaboration

However, debugging this system proved challenging. There were inconsistencies in the live tracking feed which led to frequent cost discrepancies. After countless iterations and tests, Wambugu finally resolved these issues through meticulous tuning of the OpenAI SDK calls.

Builder Notes: Working closely with Claude, they identified specific error messages returned by Protrack 365 API that caused inconsistent data readings. This required fine-tuning how their code interacted with external APIs to ensure smooth and reliable real-time data processing."

\`\` The Moment of Triumph

Finally, everything worked as intended — a system that could accurately calculate costs based on truck location in real time. Wambugu recalls the moment when he tested it live for his client: "The first delivery cost was exactly what I expected. Seeing confidence from my customer drove home how far we had come with this AI integration."

\`\` What It Does

Arise & Shine Transporters now offers a reliable, dynamic pricing model and real-time fleet visibility that eliminates the need for manual tracking.

**** Truck dashboard displaying GPS data"

\`\`

Wambugu’s Reflections

Building Arise & Shine Transporters with AI was both rewarding and challenging. Confusion initially led to many mistakes but collaboration enabled them to overcome these hurdles efficiently, resulting in a robust product that solved real-world logistics problems for Kenyan businesses.

Builder Notes: This project reinforced the importance of having strong communication channels between development teams working on integrated systems, as well as thorough testing protocols."

\`\` What I Am Proud Of

Wambugu is proud of how he and his team navigated complex integration issues to create a robust platform that delivers real value. The journey taught him valuable lessons about leveraging AI effectively for practical use cases."