Role Overview

We are looking for a curious and detail-oriented intern to help test AI-powered products for accuracy, safety, and reliable behavior.

Can you make AI invent an answer, ignore an instruction, or contradict itself? Help us find these failures, document them clearly, and verify fixes with engineers.

This role combines manual testing with basic scripting. It suits someone curious, detail-oriented, patient with repeated attempts, and willing to raise problems and suggest improvements.

???? ???'?? ??

- Take initiative in choosing what to test and exploring new failure cases.
- Turn requirements into test cases and ask questions when expected behavior is unclear.
- Test everyday conversations, unusual inputs, spelling variations, unclear requests, and longer conversations.
- Explore jailbreak prompts that attempt to make AI ignore its instructions.
- Check for invented facts, inconsistent answers, inappropriate responses, and unnecessary refusals.
- Repeat tests to understand how often failures occur and what triggers them.
- Write clear bug reports with inputs, conversation history, reproduction steps, and expected versus actual results.
- Use basic Python or JavaScript to send API requests, repeat tests, and compare responses.
- Prioritize issues, discuss improvements with engineers, and retest fixes.

???? ??'?? ??????? ???




- Curiosity about AI and how it fails.
- Patience for trial and error.
- Attention to detail and organized documentation.
- Clear communication and willingness to explain your thinking.
- Basic Python or JavaScript​: enough to read and modify simple scripts.
- Basic understanding of APIs and JSON responses.
- Ability to check confident-sounding answers against supplied facts and instructions.
- Willingness to learn from feedback.

???? ?? ????

Familiarity with Git, Postman, or curl exposure to testing or bug reporting; personal AI experiments or coding projects; or experience testing conversations in multiple languages.

Professional experience is not required. Experienced applicants are welcome if the internship's responsibilities and terms suit them.

Some testing may involve sensitive or harmful content, within company-authorized environments.

??? ?? ?????

Complete our seven-question application questionnaire.

Allow approximately 15–30 minutes; this is not a speed test. Advanced ML knowledge is not required. Partial answers with explicit reasoning are welcome.

Use of AI is not appreciated. Submit your own thinking, wording, and code. Do not use AI to generate, rewrite, or polish answers. We may ask you to explain your work later.

📌 AI QA/QC Engineer Intern (Bengaluru)
🏢 Vobble
📍 Bengaluru