Basketball footage data labeling

We're building an AI system that detects made baskets in basketball game footage. We need 2-3 freelancers to help create training data for our computer vision model.

THE WORK (two parts):

Part 1 — Find footage (~3-4 hours)
Find and download 5-7 amateur basketball game videos from YouTube. We need rec league, pickup, and high school games — NOT NBA broadcasts. Stationary camera, full court visible, phone-recorded quality. We need variety: different gyms, outdoor courts, lighting, camera angles.

Upload the videos to our Roboflow project. Roboflow splits them into frames automatically:

Part 2 — Label frames (~8-10 hours)
Draw bounding boxes around two objects in each frame:
• Ball — the basketball
• Hoop — the rim/ring only (NOT the backboard)

We provide a detailed annotation guide with visual examples, edge cases, and a quality checklist. The tool is Roboflow (browser-based, no software to install).

WHAT WE'RE LOOKING FOR:
• Experience with image annotation / data labeling (Roboflow, CVAT, Labelbox, or similar)
• Attention to detail — tight bounding boxes, correct class labels
• Reliable availability for ~12-14 hours over 1 week
• Comfortable following written instructions precisely

HOW TO APPLY:
• Mention any data labeling experience (tools used, types of projects)
• Confirm you can complete ~12-14 hours within 1 week

We start with a paid 20-frame test batch to verify quality before the full project.