Full-Stack Dev: Build 3D Gaussian Splatting Web Dashboard

Contract

|

United States

|

Zoom

Project Overview

I am looking for an experienced full-stack web developer with 3D web expertise to build a web-based 3D Gaussian Splatting (3DGS) dashboard.

The goal is to replicate the user interface and user experience of the Luma AI "Fields" dashboard (images attached), but populate it using the KIRI Engine API for processing and generation.

The application will allow users to upload video/images, process them into 3D Gaussian Splats via KIRI's API, and view the results in an interactive web viewer.

Key Features to Build

The project consists of three main views/components based on the Luma AI interface:

• Main Dashboard (Gallery View)

Grid Layout:

A clean, dark-themed responsive grid displaying user captures.

Thumbnails:

Animated or static previews of the 3D captures (similar to the provided "").

Metadata:

Display title, capture date, and status (e.g., "Processing," "Finished") on hover or as a caption.

Search/Filter:

Basic search functionality to filter captures by name.

• Creation Modal (Upload Workflow)

Drag & Drop Interface:

A "Create" modal (similar to "") allowing users to drag and drop video files (MP4/MOV) or image sets.

API Integration:

On upload, the file must be sent to the KIRI Engine API to trigger the "3DGS Scan" task.

Progress Tracking:

Real-time feedback on upload progress and processing status (polling the API for status updates).

• Interactive 3D Viewer

3DGS Rendering: A dedicated viewer page to interact with the processed Gaussian Splat.

Controls:

Orbit controls (rotate, zoom, pan) using a WebGL-based library (e.g., Luma WebGL Library, Spline, or a raw + Gaussian Splat implementation).

Share/Export:

Buttons to download the resulting .ply file or a shareable link.

Technical Requirements

Frontend:

, , or (Dark mode UI is a strict requirement).

Backend:

, Python (FastAPI/Django), or similar to handle secure API requests and manage user sessions.

API:

Deep experience with REST APIs. You must implement the KIRI Engine API specifically for the 3DGS Scan function.

3D Rendering: Experience with WebGL, , or specific 3DGS web viewers (e.g., antimatter15/splat, Luma WebGL).

Storage:

AWS S3, Firebase Storage, or similar for hosting the input videos and output model files.

Deliverables

• Source Code: Fully functional frontend and backend repositories.

• Deployment: Setup on a cloud provider (Vercel, AWS, or Netlify) with environment variables configured.

• Documentation: clear instructions on how to swap API keys and manage the database.

Contract duration of 1 to 3 months.

Mandatory skills:

API, JavaScript, Python, API Integration, React, , PHP, Web Application, Gaussian Splatting

Qualifications

  • I am looking for an experienced full-stack web developer with 3D web expertise to build a web-based 3D Gaussian Splatting (3DGS) dashboard

  • Creation Modal (Upload Workflow)

  • A "Create" modal (similar to "") allowing users to drag and drop video files (MP4/MOV) or image sets

  • , , or (Dark mode UI is a strict requirement)

  • You must implement the KIRI Engine API specifically for the 3DGS Scan function

  • 3D Rendering: Experience with WebGL, , or specific 3DGS web viewers (e.g., antimatter15/splat, Luma WebGL)

  • AWS S3, Firebase Storage, or similar for hosting the input videos and output model files

  • API, JavaScript, Python, API Integration, React, , PHP, Web Application, Gaussian Splatting

Responsibilities

  • The goal is to replicate the user interface and user experience of the Luma AI "Fields" dashboard (images attached), but populate it using the KIRI Engine API for processing and generation

  • The application will allow users to upload video/images, process them into 3D Gaussian Splats via KIRI's API, and view the results in an interactive web viewer

  • A clean, dark-themed responsive grid displaying user captures

  • Animated or static previews of the 3D captures (similar to the provided "")

  • Display title, capture date, and status (e.g., "Processing," "Finished") on hover or as a caption

  • Basic search functionality to filter captures by name

  • On upload, the file must be sent to the KIRI Engine API to trigger the "3DGS Scan" task

  • Real-time feedback on upload progress and processing status (polling the API for status updates)

  • Interactive 3D Viewer

  • 3DGS Rendering: A dedicated viewer page to interact with the processed Gaussian Splat

  • Orbit controls (rotate, zoom, pan) using a WebGL-based library (e.g., Luma WebGL Library, Spline, or a raw + Gaussian Splat implementation)

  • Buttons to download the resulting .ply file or a shareable link

  • , Python (FastAPI/Django), or similar to handle secure API requests and manage user sessions

  • Source Code: Fully functional frontend and backend repositories

  • Deployment: Setup on a cloud provider (Vercel, AWS, or Netlify) with environment variables configured

  • Documentation: clear instructions on how to swap API keys and manage the database

  • Contract duration of 1 to 3 months