Projects in Kaleem Shaheed Colony 1Projects in Kaleem Shaheed Colony 1Blizzup Agentic Bike Dealership
Project Overview
A sophisticated, full-stack AI-driven dealership platform that leverages Agentic AI to provide deep technical comparisons, transparent mathematical scoring, and automated inventory management. Built to fulfill a "Fullstack + AI Developer" assessment, the system uses a ReAct (Reason + Act) loop, allowing the AI to autonomously fetch real-world data and execute complex backend functions. (Note: The live AI agent functionality is currently disabled.)
Core Technologies & Frameworks
React, Tailwind CSS
Node.js, Express
Google Gemini AI (@google/generative-ai)
MongoDB Atlas / Mongoose
Pollinations.ai (http://Pollinations.ai) (Image Generation)
Key Features & Engineering Highlights
Agentic AI & Function Calling: Engineered a state-machine ReAct loop where the AI autonomously decides when to trigger real-time backend functions to retrieve inventory data, moving beyond traditional static prompting.
Explainable Scoring Engine: Developed a dynamic scoring system that evaluates bikes across five strict metrics (Price, Fuel Average, Engine Power, Value, Features), featuring "Thinking Accordions" that expose the AI's internal mathematical reasoning to the user.
Automated Bulk Ingestion: Built an intelligent pipeline allowing admins to ingest multiple vehicles by name; the AI automatically fetches exact technical specifications and generates high-fidelity photographic prompts.
Dynamic Image Generation & Self-Healing: Integrated Pollinations.ai (http://Pollinations.ai) to dynamically generate accurate vehicle images based on AI-classified categories (e.g., Mountain Bike vs. Superbike), alongside an automated admin utility to repair and refresh low-quality database thumbnails.