AI-Based System Design Builder

This project provides an interactive platform for building and simulating system designs using AI. Powered by NVIDIA AI Workbench, it leverages advanced AI capabilities to generate optimized, scalable system architectures.

YouTube Devpost

Prerequisites

Before running this project, make sure you have the following installed:

Installation

1. Set Up NVIDIA AI Workbench

Follow the Installation guide to set up the Workbench on your local machine or cloud environment.

Start the local server in Nvidia AI Workbench

NVIDIA AI Workbench Setup

2. Select Clone Project Option

Use this repository to clone:

https://github.com/pvbgeek/NvidiaDellHackathon-AI-SystemDesignBuilder Clone Project Screen

3. Project Build Initiation

When you clone the project, The project will first appear in your main window with a "BUILD REQUIRED" status. Shortly after, the status will automatically change to "BUILDING" as the system begins the build process.

Build Required Status Building Status

4. Building Stage

After clicking on the project, it will enter the building stage. Please be patient, as the build process may take between 2 to 5 minutes to complete.

Building Stage

5. Monitor Build Progress

By selecting the output tab at the bottom left of the window, you can monitor the building process in real-time. Once the build finishes successfully, you'll see "build completed" in the output, and the status will change to "build ready" in the bottom right corner.

Build Progress

6. Launch Application

After the build process completes, a green button labeled "Open AI-System Design Builder" will appear at the top of the window. Click this button to launch the application.

Launch Application

7. Application Deployment

After launching the application, the Terminal window under the "Application" category will display the output log. Simultaneously, a new browser tab will automatically open, showing your application running on localhost - indicating successful deployment.

Terminal Output Browser View

8. Creating Graphs

You can create graphs in two ways:

Graph Creation Interface

9. Processing Queries

After submitting your query, a processing message will appear. The design generation time varies between 30-45 seconds, depending on how complex your request is.

Query Processing

10. Final Design

The design has been completed and is now ready to use. You can freely customize and modify it according to your needs.

Final Design

How we built it

We developed the AI System Design Builder using a combination of cutting-edge technologies and platforms:

NVIDIA AI Workbench

NVIDIA AI Workbench

Python

Python

Flask

Flask

HTML

HTML and CSS

JavaScript

JavaScript

GenAI

GenAI

Docker

Docker

Challenges we ran into

  1. Integrating AI Capabilities: Incorporating GenAI for accurate system design generation.
  2. Balancing Flexibility and Constraints: Ensuring versatility while maintaining a predefined component set.
  3. Optimizing Performance: Achieving quick and efficient AI-generated designs.

Accomplishments that we're proud of

  1. Intuitive User Interface: Creating an accessible system design tool.
  2. AI-Powered Design Generation: Successfully implementing AI for system architecture creation.
  3. NVIDIA AI Workbench Integration: Leveraging advanced AI capabilities.

What we learned

  1. AI in System Design: Applying AI to traditionally human-centric tasks.
  2. Working with NVIDIA AI Workbench: Harnessing its power for AI-driven development.
  3. Balancing User Input and AI Assistance: Creating a powerful hybrid approach.

What's next for AI System Design Builder

  1. Expanded Component Library
  2. Enhanced AI Capabilities
  3. Collaboration Features
  4. Performance Analysis
  5. Integration with Development Tools

Meet the Team

Parth Bhalerao

Parth Bhalerao

Worked on front-end and back-end JS development, creating a seamless user experience and robust application logic.

LinkedIn Profile
Atharva Weginwar

Atharva Weginwar

Developed the GenAI feature, enabling AI-powered system design generation based on user requirements.

LinkedIn Profile
Ashutosh Talwalkar

Ashutosh Talwalkar

Focused on front-end development, crafting an intuitive and visually appealing user interface.

LinkedIn Profile
Yugandhar Desai

Yugandhar Desai

Led the deployment of all features on NVIDIA AI Workbench, ensuring smooth integration and performance.

LinkedIn Profile