RoleGPSThe free map to your AI job
All capabilities · Annotation, quality & human feedback

Work in annotation tools (Label Studio / CVAT / Labelbox)

Set up projects, hotkeys, bounding boxes/segmentation/NER, export formats.

~6 focused hoursbeginner
Explore 3 tools for this project
Market relevance

Which roles ask for this — and how often

Share of job postings in India, per role, that name this capability.

AI Data Annotator / Labeling QA13%Computer Vision Engineer8%
What employers mean

You should be able to…

  1. Set up a new labeling project with the right configuration (bounding box, NER, classification, segmentation)
  2. Use keyboard hotkeys to hit realistic throughput targets, not just mouse clicks
  3. Export labeled data in the format a downstream model actually needs (COCO/YOLO/CoNLL/JSON)
  4. Configure multi-annotator review/consensus workflows inside the tool
  5. Import pre-annotations or model predictions to speed up human review
  6. Troubleshoot common tool issues (export mismatches, project config errors)
  7. Move between tools quickly since different clients standardize on different platforms

Needs first: Label data accurately against guidelines

Learn — free, link-checked

The few resources that matter

Read · beginner · 25 min · docs.labelbox.com

Platform overview

Covers project setup and export formats in Labelbox, the third major annotation platform named in Indian AI job listings. — Labelbox
Read · beginner · 30 min · labelstud.io

Get started with Label Studio

Official quickstart for installing Label Studio and configuring your first labeling project with hotkeys. — HumanSignal / Label Studio
Read · beginner · 30 min · docs.cvat.ai

Getting started

Official guide to bounding boxes, segmentation, and export formats in CVAT, the most common open-source CV labeling tool named in job posts. — CVAT.ai
Read · intermediate · 25 min · labelstud.io

Set up task agreement

Shows how to configure and read inter-annotator agreement directly inside the tool you'll be evaluated in. — HumanSignal / Label Studio
Tools for practice

Choose a tool for the job

Start with one tool for each part of your project. You don’t need to learn them all.

Go to the practice brief

3 tools to explore

Label Studio

Data · Test

Label examples, compare annotations and export a dataset for review or evaluation.

Visit toolGetting started

Labelbox

Data · Test

Organise a labelling project and review annotations against a consistent rubric.

Visit toolGetting started

Practices & references

  • COCO, YOLO and JSON export formats
  • Annotation hotkeys

Selected for practice by RoleGPS. No paid placements or affiliate links.
Access can change; check the details before starting.

Think your tool is a good fit? Let us know

Practice

Bounding-box labeling project in Label Studio, exported to COCO

Set up a Label Studio (or CVAT) project to bounding-box-label 100 street-scene images from the public COCO val2017 set for three object classes. Configure the label schema and keyboard shortcuts, label the whole batch working at speed, export to COCO JSON, and write a validation script that reloads the export and confirms it is well-formed. Document the label config so someone else could stand up the identical project from your README.

Start from

COCO val2017 images — 5k public photos, free to download without signup; pick 100 street scenes

Milestones
  1. Install Label Studio, define the 3-class label config and the hotkeys · ~1.5h
  2. Label the 100 images in hotkey-driven passes · ~1.5h
  3. Export to COCO JSON and write the script that validates the export loads cleanly · ~1h
  4. Document the label config and record the labeling screen capture · ~1h
Done when
  • Project is configured with the correct label schema for 3 classes and keyboard shortcuts are used (documented in README)
  • All 100 images are labeled and exported in valid COCO JSON that loads without error
  • README documents the project config (label config XML/JSON) so someone else could reproduce the setup
  • At least one screenshot/recording of the labeling interface with hotkeys in use
Prove it

Evidence a recruiter can check

  • The exported COCO JSON for all 100 images, alongside a validation script that reloads it and prints per-class box counts
  • The Label Studio label config (XML) and hotkey map, complete enough to recreate the project from scratch
  • A screen recording of hotkey-driven labeling with your actual images-per-minute rate stated
  • A note on the labeling decisions you had to make — occluded objects, truncated boxes, what counts as a separate instance
Signal it

Labelled 100 street-scene images with bounding boxes in Label Studio using a documented 3-class schema and hotkeys — exported valid COCO JSON and wrote a validator that checks the export loads cleanly.

Interview

Questions you'll get asked

  1. Walk me through setting up a new bounding-box project in CVAT or Label Studio from scratch.
  2. How do you configure a project so two annotators label the same items for agreement checking?
  3. What export format would you use for a YOLO object-detection model, and why does the format matter?
  4. How would you speed up labeling for a project with 50,000 images using pre-annotations?
  5. What hotkeys or shortcuts do you rely on to hit a daily throughput target?
  6. Describe a time a tool's export didn't match what the ML team needed — how did you fix it?
See where you stand for AI Data Annotator / Labeling QA