Website:
petyosa.com
Job details:
๐ฅ๐ผ๐น๐ฒ ๐ฆ๐๐บ๐บ๐ฎ๐ฟ๐:
Build and run PetYosa's products end to end by directing AI coding agents โ and prove every change works for a real customer before calling it done.
๐๐ฒ๐๐ฐ๐ฟ๐ถ๐ฝ๐๐ถ๐ผ๐ป:
PetYosa is a pet-care platform in India: an app for pet parents, and software for the vets, groomers, boarding homes and pet shops who serve them.
The unusual part. PetYosa's founder is not an engineer. In August 2026 alone, the founder made about 1,400 code changes to PetYosa's apps and servers โ plus nearly 500 updates to its written operations playbooks โ by directing AI coding agents through a strict build method.
We are hiring engineers who can run that same method and bring real engineering judgement on top of it. This is a full-time, office-based role in Bangalore.
The method you'll work in
Every change goes: build โ smoke test โ adversarial review โ fix โ commit โ push โ merge โ deploy from the merged commit โ prove a customer can actually see it. A bug is never fixed for just one person โ you find and fix every place the same problem exists.
How we hire โ no whiteboards, no puzzles
โข Short application (10 minutes). Your CV and answers are first checked automatically.
โข 6 online proctored tests โ no AI, external monitor or screen sharing allowed, camera and microphone on (4 hours in total)
โข One/Two conversations
All six tests are scored automatically. Test 1 usually reaches you within a few minutes of applying, though occasionally it takes longer. Each later test is emailed as soon as you pass the one before. Please check your spam folder too.
๐๐ฒ๐ ๐ฅ๐ฒ๐๐ฝ๐ผ๐ป๐๐ถ๐ฏ๐ถ๐น๐ถ๐๐ถ๐ฒ๐:
โข Own features end to end across the PetYosa consumer app, the partner platform (petyosasparsh.com), the admin console and the APIs behind them.
โข Testing what you are going to ship and what you shipped
โข Build with Claude Code and Codex as your primary tools, running several agents in different roles โ one builds, one reviews, one attacks.
โข Run the method on every change, and close only on proof a customer can see โ never on "merged" or "deployed".
โข When a bug is reported, find the whole class of it across the platform and fix it everywhere.
โข Keep the written playbooks current, in the same change as the work.
โข Write short, plain-English updates the founder can act on.
๐ฅ๐ฒ๐พ๐๐ถ๐ฟ๐ฒ๐ฑ ๐ฆ๐ธ๐ถ๐น๐น๐:
โข You already build mostly by directing AI agents (Claude Code, Codex CLI) โ and can show us something real you shipped that way.
โข Your reflex after "it works" is "now let me break it".
โข You can tell when an AI is confidently wrong, and you check before you trust.
โข You read TypeScript, React, Node, MongoDB/PostgreSQL and Docker well enough to judge what an agent wrote.
โข When someone says your fix didn't work, you verify before you defend.
โข You write plain English that a non-technical founder can act on.
What doesn't matter: your college, big-company names, or algorithm puzzles.
๐ฃ๐ฟ๐ฒ๐ณ๐ฒ๐ฟ๐ฟ๐ฒ๐ฑ ๐ฆ๐ธ๐ถ๐น๐น๐:
โข Linux servers / hypervisors
โข Playwright / end-to-end testing
โข Mobile builds (Capacitor / Android)
โข Razorpay / WhatsApp Business API
๐ง๐ผ๐ผ๐น๐ / ๐ง๐ฒ๐ฐ๐ต ๐ฆ๐๐ฎ๐ฐ๐ธ:
- Claude Code / Codex CLI
- TypeScript, React, Node / Bun (Hono, Express)
- MongoDB, PostgreSQL / Supabase
- Capacitor (iOS / Android)
- Docker, Linux, git
๐๐ง๐: INR 20,00,000 - 24,00,000
๐ฉ๐ฎ๐ฟ๐ถ๐ฎ๐ฏ๐น๐ฒ ๐ฃ๐ฎ๐: Excellent Long-Term Incentive Plan (LTIP) for great candidates
๐ช๐ต๐ ๐๐ผ๐ถ๐ป ๐จ๐:
The method is already written down, so you'll spend your time shipping โ not guessing how we work โ and you'll build with AI agents every day.
๐๐ฝ๐ฝ๐น๐ ๐๐ฒ๐ฟ๐ฒ: https://petyosa.hiresense.co.in/apply/63742313d7a9c370489cb1f4d71befbb
Click on Apply to know more.