SDE-2 (Backend) · Customer Growth · Urban Company

Gurugram || Work from office || 2.5–4 years of experience

About Urban Company

Urban Company is India's largest home-services platform, and also operates in the UAE, KSA and Singapore. Over 8 million customers a year use our app to book everything from a salon visit or a jet-spray AC service to UC's own Native RO water purifier, delivered at home by trained professionals.

About the role

Join the Customer Growth team, which builds the tech for:

1. Customer acquisition and retention: plans, bundles, and AI voice agents that help customers who dropped off mid-booking finish it
2. Booking journey, catalog and pricing: search, checkout, an in-app voice concierge, and the engine that prices every booking
3. Marketing, personalisation and recommendations

When this team ships, bookings and revenue move within weeks.

How we build

Every UC engineer uses AI coding agents daily, and Claude Code is standard. We hire engineers who direct agents well and own every line that ships, whoever wrote it. That takes solid fundamentals: you can't review an agent's idempotent retry without knowing what a correct one looks like.

What you'll do

- Own outcomes, not tickets. Take a problem like "customers abandon checkout", then ship, measure and iterate.
- Build and run high-traffic backend services on Node.js/TypeScript, MongoDB, MySQL, Redis and event-driven architecture.
- Work through AI agents: write the spec, delegate, review and test. Improve the team's shared agent workflows.
- Write design docs that weigh the trade-offs and make a call.
- Own production for what you build: latency, availability, observability and incidents.
- Run growth experiments with product, design, marketing and operations.

What we expect from you

- Experience. 2.5–4 years of backend engineering, in any language.



At least experience at a product company or startup, building and running your own product's systems.
- Production at scale. In the last 18 months, backend code that serves real traffic, and you can put numbers on it: users, requests per second, orders a day, data volume.
- End-to-end ownership. A system or feature that was yours from design through production to the metric it moved, including fixing it when it broke.
- Managing AI agents. Claude Code, Cursor, Codex or similar are how you work every day, not an occasional autocomplete. You:
- write a plan or spec first, and break big tasks into pieces an agent can do reliably
- give agents the context they need, such as conventions, docs and examples
- run several agents in parallel and keep each one on track
- check output with tests and by reading the diff, and know the mistakes agents tend to make in your codebase
- Self-drive. You've shipped something nobody assigned you because you saw the problem first, and you can tell us what changed as a result.
- Problem solving. You turn a vague goal into a measurable one, fix the root cause rather than the symptom, and cut scope to ship. Solid data structures and algorithms.
- Fundamentals. Async and event-driven design, queues, API design, data modelling and indexing in SQL and NoSQL, caching, idempotency and retries, and what happens when a dependency fails.

Good to have: LLM-powered features or agents shipped to production, evals

Why join

- Scale. Your code prices, ranks and books services for over 8 million customers a year.
- AI-first in practice. You'll use frontier coding agents every day and build AI products that customers talk to.
- Ownership early. SDE-2s own services and metrics, not just a slice of someone else's project.
- A high bar in a small team. You'll work with careful reviewers, move quick and get feedback in days.

📌 SDE - II - Growth (Gurugram)
🏢 Urban
📍 Gurugram