Junior Business Analyst (WFH)

About the job:
Key responsibilities:

1. Label people, products, hands, registers and actions in store video and images using our internal tools and tools like CVAT or Label Studio.

2. Draw bounding boxes, polygons and zones, and tag clips with event types (theft, void, no-sale, idle staff and others).

3. Write short, precise text descriptions of events for vision-language model training.

4. Follow labelling guidelines exactly, and raise unclear or edge cases with the team.

5. Select and pull clips and frames from recorded footage for each new dataset.

6. Clean datasets: remove duplicates, blurry or irrelevant frames, and mislabeled items.

7. Match video events to point-of-sale transactions by store, camera and timestamp.

8. Balance datasets across stores, camera angles, lighting and times of day.

9. Keep dataset metadata, versions and train/test splits organised and documented.

10. Review other annotators' work and keep labels consistent.

11. Review AI predictions and alerts, and mark each one correct, incorrect or unclear, with a reason.

12. Build and maintain 'ground truth' test sets that new models are scored against.

13. Track false alarms and missed events, and sort errors by cause (camera angle, lighting, occlusion, store layout).

14. Summarise results in Google Sheets or Excel: counts, accuracy, and examples of each error type.

15. Test new models and features before release, and report what got better or worse.

16. Hit agreed targets for label accuracy and throughput.

17. Keep customer video and store data confidential at all times.

18. Sign NDA on the process and customer privacy

Who can apply:

Only those candidates can apply who:
- have minimum 1 years of experience

Salary:
₹ 2,00,000 - 2,40,000 /year

Experience:
1 year(s)

Deadline:
2026-11-12 23:59:59

Other perks:
5 days a week

Skills required:
MS-Office, English Proficiency (Written), Data Analysis, Data Preparation, Data Extraction and Data Annotation

Other Requirements:
- 1. Computer Science Graduates
- 2. Data Engineering
- 3. Artificial Intelligence

About Company:
Jinn Labs builds vision AI for convenience stores. Our systems run on cameras and point-of-sale data in retail stores across the United States, flagging theft, cash-register issues and staff productivity in real time. AIM is our human-in-the-loop team. AI spots possible incidents; AIM analysts confirm what is real, alert store owners within seconds, and turn every decision into training data that makes the models better. You will sit at the point where people and AI meet.