We’re seeking a highly organized and technically proficient contractor to help us build an AI-powered image tagging and search system for client folders stored in Dropbox. The goal is to enable easy retrieval of visual assets (e.g., patios, fire pits) by automatically tagging and categorizing them, while also establishing a repeatable, scalable workflow for future uploads.

ideally, we are not creating from scratch but rather integrating a tool that can "read" existing catalog of images on dropbox and allow us to search by what is in the image (without the need for tagging).

This role involves building a system that supports:

Automated tagging and categorization (e.g., “fire pit,” “patio,” “pergola,” “outdoor kitchen”).

Clear handling of “Do Not Use” (DNU) images, including flagging, segregating by client, and cross-referencing to ensure excluded assets don’t appear in deliverables.

Designer checkpoints (e.g., tagging images as “approved” or “not approved” for use).

Consistent naming conventions and comma-separated multi-feature tags.

Easy retrieval for case studies, marketing projects, and client-facing deliverables.

Ongoing tagging for new client uploads and the ability to label new/unreviewed images.

Key Responsibilities

Evaluate and recommend the best AI-based tool (Clarifai, AWS Rekognition, or Google Vision API) for landscaping/outdoor imagery.

Configure AI models to detect and tag relevant features (fire pits, patios, pergolas, grills, etc.).

Build a no-code/low-code workflow (Zapier, Make, or light scripting) to:

Connect Dropbox with the AI tool.

Apply labels/tags consistently.

Move or organize images into appropriate folders.

Export metadata (tags, DNU flags, approval status) into a spreadsheet or database.

Establish a DNU (Do Not Use) management system, including folder structures, tagging, and AI-based cross-checks.

Implement a “designer first-touch” checkpoint to tag/approve usable vs. non-usable images.

Document naming conventions and multi-feature tagging standards.

Set up automated labeling for new uploads (e.g., “new,” “unreviewed”).

Provide documentation and basic training so the internal team can manage and replicate the workflow.

(Optional) Assist with custom model training to improve detection accuracy over time.

Qualifications

Proven experience with one or more: Clarifai, AWS Rekognition, Google Vision API.

Experience integrating cloud tools with Dropbox using Zapier, Make, or Python.

Strong understanding of image recognition, classification, and automation workflows.

Bonus: Familiarity with DAM (Digital Asset Management) systems and best practices for large image libraries.

Detail-oriented, with a focus on documentation, scalability, and repeatability.