What you will own
- Release quality. You have the final call on whether a build ships. Not advisory. Actual veto.
- The automated test suite. Build it from close to zero. API, web, mobile, and edge pipeline coverage.
- Hardware-in-the-loop testing. Our bugs live where models, cameras, GPUs, and flaky networks meet. Set up a rig that reproduces field conditions on a desk.
- Non-deterministic system validation. CV detections and voice agent responses do not pass or fail on string equality. Design accuracy thresholds, golden datasets, drift checks, and regression baselines that catch a 3% mAP drop before a client does.
- Field failure loop. Every production incident becomes a permanent test case within 48 hours.
- Quality metrics that mean something. Escaped defect rate, mean time to detect, flake rate, coverage on critical paths. You publish them weekly.
How you will work
You are expected to use AI agents as a force multiplier, not as an occasional trick.
- Use coding agents (Claude Code, Cursor, or equivalent) to generate and maintain test suites at a volume a manual writer cannot match
- Build agent-driven exploratory testing that navigates our web and mobile apps and reports anomalies
- Write specs and prompts precise enough that agents produce test code you would have written yourself
- Automate your own reporting so you are not spending Fridays making status updates
If your answer to "how would you test 400 API endpoints in a week" involves writing 400 tests by hand, this is not the role.
Must have
- 2+ years in test automation, not manual QA with some Selenium exposure
- Strong Python. This is our stack. JavaScript/TypeScript for frontend testing
- Playwright or Cypress, pytest, API testing, CI pipelines (GitHub Actions)
- Linux comfort. Docker. You can SSH into a device and read logs without help
- Daily working use of AI coding agents, with examples you can show us
- Ability to reproduce a bug from a vague client message and a screenshot
Strong plus
- Tested ML or CV systems. You know why "the model is 94% accurate" is not a test result
- Jetson, embedded Linux, or any experience testing on real hardware
- Load and performance testing (k6, Locust)
- Mobile automation (Appium, Detox) for React Native or Flutter apps
What we mean by ownership
- You find the gap and close it without a ticket being assigned
- When something breaks in production in another country, you are not waiting to be asked
- You push back on engineers, on the CTO, and on CEO when quality is being traded away
- You do not say "that is a dev problem"
What this job is not
- Not a QA process role. No test plan templates, no sign-off theatre
- Not a stepping stone to a dev role. If you want to be a backend engineer, apply for that instead
- Not comfortable. Bootstrapped company, real deadlines, clients who notice
Hiring process
- Application with a short note: hardest bug you found, and how
- 30 min call with the interview team
- Take-home assessment: we hand you a small buggy service, you find what is wrong and automate the regression
- Technical discussion of your submission
- Offer
Pay: ₹320,000.00 - ₹500,000.00 per year
Benefits:
- Flexible schedule
- Leave encashment
- Work from home
Ability to commute/relocate:
- Noida, Uttar Pradesh (Noida): Reliably commute or planning to relocate before starting work (Required)
Application Question(s):
- Can you join Immediately? If not what is your official notice period?
- What are your expectations - CTC?
Experience:
- QA/QC: 2 years (Required)
Language:
Work Location: Hybrid remote in Noida, Uttar Pradesh (Noida)
