QA Engineer (Manual + Automation)

We're hiring on behalf of our client - an AI-native Startup.

Overpath mission Overpath builds Molly, an AI copilot for Enterprise revenue teams. Molly joins sales calls, reads the deal history across CRM, email and meetings, and delivers insights, tasks and coaching to reps and managers inside Slack, Google Chat and our web app.

We ship to production every week, and we work with live design partners who feel every release. Today, the founders and engineers do the pre-release testing themselves. That stops with you. You will be Overpath's first dedicated QA engineer: the person who makes the staging-to-production jump solid, and who writes the quality playbook this team scales on.

What you'll do Own release quality

• Run the regression pass on staging before every production release: Molly's chat flows in Slack, Google Chat and the web app, onboarding, notifications and task management
• Verify our integrations end to end: HubSpot and Salesforce syncs, calendar events, meeting-recorder pipelines including reconnects, permission changes and partial data
• Hunt edge cases through exploratory testing, and turn every bug you find into a reproducible, written test case

Test AI, not just software

• Judge whether Molly's LLM-generated responses are correct, grounded in the right deal data and safe to ship - AI output does not crash, it convinces, and that makes correctness your job
• Help define what "pass" means for non-deterministic features, working with the team's eval harness rather than exact-match assertions
• Stress-test our agentic workflows (multi-step reasoning over deals, transcripts and CRM data) a from staging to production

Build the automation suite

• Start manual to find the edge cases, then convert the repetitive scenarios into automated test cases that run on every PR
• Document test workflows so precisely that the dev team (and their AI coding tools) can turn them into automation scripts directly - your writing becomes our test infrastructure
• Help enforce our Definition of Done: dev done vs released vs QA'd

Write the playbook

• There is no QA department here and no existing process to inherit; the test-case library, the release checklist and the quality bar are yours to define
• What you build becomes how Overpath does quality as the team grows

Your profile Testing craft

• 2 to 4 years of hands-on QA in a SaaS product, with real depth in manual, exploratory and regression testing - this is a dedicated testing role, not a development role with testing on the side
• Strong test-case design and documentation: workflows another person can execute exactly, first time
• Experience testing third-party integrations, webhooks and multi-environment setups (staging vs production)
• Working knowledge of automation (Playwright, Cypress or Selenium) and API testing (Postman or similar), with enough JavaScript/TypeScript to read and adapt scripts

AI-native mindset

• You already use AI tools daily to write test cases, generate data and speed up your own work - as a core part of your workflow, not a novelty
• You are curious about how LLM products fail, and you think in score trends across releases, not one-off pass/fail

Builder mentality

• Energised by being the first: no playbook to follow, so you write it
• Comfortable with weekly releases, shifting scope and direct communication in a small team
• Proactive by default: you flag the risk before anyone asks about it

You might be coming from • A QA or test engineer role at a SaaS startup where you were one of very few testers

• A services or product company QA team, wanting ownership instead of ticket queues
• A manual testing background, actively building automation skills with AI assistance
• Any role where you tested a chatbot, agent or other LLM-powered feature and got hooked on the problem

Why us • Be the first QA hire at an AI-native startup at an inflection point: your playbook becomes Overpath's playbook

• Work directly with the CTO and a small senior engineering team, with no layers in between
• Test a genuinely hard surface: agentic AI over real revenue data, across chat platforms, CRMs and meeting recorders
• Remote, India-based