MH.
Back to projects
Full-Stack

Smart Agriculture

Practical AI for agriculture: disease detection workflow built for fast field decisions.

React.jsNode.jsExpress.jsGemini APIMongoDBJWT
AI IntegrationComputer Vision WorkflowAuthenticationProduct UX
Live ReviewGitHub
Smart Agriculture cover

Overview

Smart Agriculture helps users upload crop images, receive AI-powered disease analysis, and track previous predictions. The product combines a simple frontend workflow with secure backend services for repeated use.

Problem

Farmers and operators often need diagnosis quickly, but expert access can be delayed and decision quality drops without a reliable first-pass signal.

Solution

I implemented a full-stack prediction pipeline with image validation, Gemini integration, and authenticated history so users can compare outcomes and make faster decisions.

Architecture

smart-agriculture

Frontend

React.js

Backend

Node.jsExpress.js

Database

MongoDB

AI Layer

Gemini API

External Services

JWT

Impact

  • Reduced friction from image upload to prediction result in a single guided flow.
  • Improved trust with authenticated history and stable API error/loading states.
  • Connected AI inference to practical user actions instead of a standalone demo model.

Features

  • Leaf image upload with validation and processing safeguards
  • Gemini-based disease prediction service integration
  • JWT-secured user accounts and protected endpoints
  • Prediction history per user for comparison and follow-up
  • Clear loading, empty, and failure states in the UI
Engineering Details

System Overview

Smart Agriculture helps users upload crop images, receive AI-powered disease analysis, and track previous predictions. The product combines a simple frontend workflow with secure backend services for repeated use.

Tech Stack

React.jsNode.jsExpress.jsGemini APIMongoDBJWT

Challenges

Farmers and operators often need diagnosis quickly, but expert access can be delayed and decision quality drops without a reliable first-pass signal.

Architecture

smart-agriculture

Frontend

React.js

Backend

Node.jsExpress.js

Database

MongoDB

AI Layer

Gemini API

External Services

JWT

Key Decisions

I implemented a full-stack prediction pipeline with image validation, Gemini integration, and authenticated history so users can compare outcomes and make faster decisions.

Outcomes

  • Reduced friction from image upload to prediction result in a single guided flow.
  • Improved trust with authenticated history and stable API error/loading states.
  • Connected AI inference to practical user actions instead of a standalone demo model.

Screenshots