Software Engineer

About the job

Job Overview
We are hiring engineers who build with agents, not just with editors.

Mumzworld is the largest mother, baby and child e-commerce platform in the Middle East — millions of customers, a deep catalogue, and a stack that has to stay fast, correct and cheap at scale. We are rebuilding how our engineering team works around AI: agents write, review and test alongside us, and the engineers who compound fastest are the ones who know how to direct them.

This is a backend-heavy full stack role. You will own services end to end — schema to API to the React surface that consumes it — and you will be expected to ship at a pace that only makes sense if you are genuinely fluent with modern AI tooling.

Key Responsibilities

• Own backend services end to end. Design, build and operate NestJS/TypeScript services on PostgreSQL — data modelling, APIs, background jobs, caching, observability, on-call for what you ship.
• Close the loop on the frontend. Build the React interfaces that consume your APIs. You should be comfortable shipping a feature all the way to the customer without a handoff.
• Work at real e-commerce scale. Catalogue, cart, checkout, pricing, promotions, inventory, order orchestration — systems where correctness and latency both matter, in two languages and multiple markets.
• Build with agents by default. Use Claude Code and similar tooling as a core part of your workflow: planning, implementation, refactors, test generation, migrations, incident triage.
• Make the team faster, not just yourself. Author agent skills, subagents and workflows that other engineers reuse. Wire up MCP servers to our internal systems. Turn tribal knowledge into tooling.
• Ship AI into the product. Build LLM-powered features — search and discovery, content generation, support automation, merchandising assistance — with proper evals, guardrails and cost/latency budgets.
• Raise the bar on quality. Strong code review, meaningful tests, and a healthy scepticism about agent-generated code that looks right.
  WHAT "AI-NATIVE" MEANS HERENot "I use Copilot sometimes." Concretely, we expect you to:
• Drive a coding agent through a non-trivial change — scoping context, breaking work down, iterating on failures — rather than accepting the first output.
• Write agent skills and reusable prompts/workflows that encode how work should be done in a specific codebase or domain.
• Compose multi-step and multi-agent workflows for things like large refactors, test backfills, dependency upgrades and codebase audits.
• Understand MCP (Model Context Protocol) — and be able to build or integrate a server that gives agents safe access to a real system.
• Evaluate LLM output properly: eval sets, regression checks, cost and latency tracking. If you cannot measure it, you cannot ship it.
• Know where agents fail — and review their output like you would a talented but overconfident junior engineer.
  

Requirments

• 3+ years building and running production software, with backend as your centre of gravity.
• Strong TypeScript/Node.js; production experience with NestJS or a comparable structured framework.
• Solid PostgreSQL — schema design, indexing, query tuning, migrations, transactions. You know why your query is slow.
• Working React proficiency — you can build and ship a clean, performant UI.
• Real experience with REST/GraphQL API design, async patterns (queues/events), caching and integration with third-party systems.
• Comfort with cloud infrastructure, CI/CD, Docker and modern observability tooling.
• Demonstrable, daily use of AI coding agents — and an opinion about how to use them well.
• Clear written communication. In an AI-native team, the quality of your writing is the quality of your engineering.

BONUS POINTS

• E-commerce, marketplace, retail or logistics systems experience.
• Built and shipped an LLM feature into production — including the unglamorous parts (evals, fallbacks, cost control).
• Published MCP servers, agent skills, plugins or open-source tooling.
• Experience with high-traffic systems, search relevance, or recommendation/personalisation.
• Arabic/English bilingual product experience.

OUR STACK

TypeScript · Node.js · NestJS · React · PostgreSQL · Redis · REST & GraphQL · Docker · AWS · GitHub Actions · Claude Code and agentic tooling throughout.

WHY MUMZWORLD

• The region’s category leader — real scale, real customers, real problems.
• An engineering org actively investing in AI-native ways of working, not just talking about them.
• Small teams, high ownership, short path from idea to production.
• Competitive package, health cover, and a genuine say in the tools you use.