QA Engineer — Hardware-Software Systems

Company: SwitchOn (Abee Research Pvt Ltd)

Location: Bengaluru (on-site, with occasional visits to customer factories)

About SwitchOn

SwitchOn is a Series A deep-tech startup eliminating defects in manufacturing with AI-powered vision inspection. Our systems run on factory lines at Unilever, ITC, P&G, Godrej, SKF, Bajaj, Brakes India and other leading manufacturers — inspecting products in real time, 24/7, where a missed defect can mean a recalled batch.

About the role

Our stack runs end to end: cameras and lighting on the line, edge hardware, firmware, ML inference, backend services, and a customer-facing dashboard. A bug anywhere in that chain shows up on a factory floor. We're looking for a QA engineer who can understand the whole system, build the automation that catches failures before our customers do, and still get hands-on to break things a script wouldn't think of.

You'll be our first dedicated QA hire — you'll shape how testing works here, not inherit a process.

What you'll do

Learn the full hardware-to-dashboard stack and map where failures originate — camera drops, network flakiness, model drift, UI regressions, config errors.

Build and own the test automation framework from the ground up: UI (dashboard), API (backend services), and integration tests across the edge-to-cloud pipeline.

Wire automated suites into CI/CD (GitHub Actions) so every PR and release candidate runs regression automatically, with clear pass/fail gating.

Set up hardware-in-the-loop test rigs — camera, lighting, edge device — so firmware and inference changes are validated against real captures, not just mocks.

Build synthetic and recorded test datasets (image sets, sensor streams, edge-case product variants) that can be replayed to reproduce field failures deterministically.

Simulate factory conditions — poor connectivity, power cycles, hardware faults — and verify the system degrades gracefully.

Work with developers from spec stage to design test plans for new features, and reproduce and triage issues reported from the field.

Create testing guidelines and release criteria the dev team follows, and own the go/no-go call on releases.

Track and drive automation coverage, flaky-test rates, and time-to-detect regressions as team metrics.

What we're looking for

3+ years in test automation for hardware-software systems — IoT, robotics, industrial automation, med-tech or similar — with at least one framework you built or substantially owned.

Strong Python; production experience with Pytest plus at least one of Playwright, Selenium or Cypress for UI, and requests/httpx or Postman/Newman for APIs.

Experience integrating test suites into CI pipelines and keeping them fast and reliable at scale.

Comfortable with Linux, Docker, shell scripting, and reading logs and code to isolate root cause.

Understanding of test design for non-deterministic systems — you know how to test an ML model's output without asserting exact values.

A customer-first instinct: you think in terms of what the plant operator sees at 2 a.m., not just whether the ticket passes.

Clear written communication — bug reports and test docs that developers act on without a follow-up meeting.

Nice to have

Experience testing computer-vision pipelines or validating ML models (accuracy regression, drift detection, golden datasets).

Hardware-in-the-loop or embedded test experience (serial/CAN/GPIO automation, power-cycle rigs).

Performance and load testing tools (Locust, k6) on real-time or streaming data.

Familiarity with PLCs, OPC-UA or Modbus, or shop-floor environments.

Why join us

Ownership from day one: define QA at a company where your work directly decides whether a production line ships good product.

Deep-tech problems most QA roles never touch — hardware, ML and software failing in combination.

Backed by Pi Ventures, Axilor, Chennai Angels and other leading deep-tech investors in India and Singapore.

A clear path into a QA lead role as the team grows.