Senior Full-Stack Engineer

CrowdReply; Pakistan; Remote; PKR 7,722,000 - 9,267,000; Full time

Benefits offered by this job

Remote work in Pakistan

Work with founders

Real ownership

Job summary

CrowdReply, Remote in Pakistan, is seeking a senior backend-leaning engineer who builds with Claude Code or Cursor daily and owns what ships. You’ll think before the agent writes a line and own production features.

You’ll work on real systems with a MERN-like stack and AI tooling, collaborating with a global team across Asia, Europe and the US. This full-time remote role offers ownership from Pakistan.

Qualifications

• 7+ years as a software engineer with backend strength: APIs, databases and queues.
• Experience designing scalable data models and fixing slow queries.
• Owned production features for 1+ year: shipping, scaling, decisions.
• Hands-on with Claude Code, Cursor or Codex and proper setup: rules, MCP servers.
• Plan-first mindset with the ability to evaluate agent-driven approaches.

Responsibilities

• Own features end to end from rough brief to shipped production.
• Plan before you build: data model, failure cases, what to build.
• Work on messy systems: AI models, rate-limited APIs, queues and billing.
• Research unfamiliar tech and return with a concrete plan.

Skills

APIs

Databases

Queues

Infrastructure

Tools

Claude Code

Cursor

Codex

Job description

CrowdReply · Remote, from anywhere in Pakistan · Full-time · $2,500–$3,000 a month

CrowdReply gets brands recommended when buyers ask ChatGPT, Perplexity or Gemini what to use. Our product runs thousands of buyer questions through those models every day and acts on the answers. We're a startup founded in 2025 and growing fast: 10,000+ brands trust CrowdReply, and we're 20+ people across Asia, Europe and the US.

You plan it. The agents build it. You own it.

We're hiring a senior, back-end leaning engineer in Pakistan who builds with Claude Code or Cursor every day, thinks before the agent writes a line and owns what ships.

AI first is how we work

The founders run this loop every day for code, design and copy. You'd be expected to run it too, and to make it better.

• Brief it properly. Give the agent the whole picture: the goal, who it is for and what must not happen. Then ask it to challenge the brief before anything gets built.
• Plan before code. Get options, weigh them and decide. The plan is yours. The agent works from it, not the other way round.
• Run agents in parallel. One builds, one researches, one reviews, each on its own task, so a day of work fits in an afternoon.
• Review it like a sceptic. Check the real result, not what the agent says it did. Fix what is wrong, then write down what you decided.

What you work with: Claude Code, Cursor, a CLAUDE.md in every project, shared team skills, our own MCP server and a company context bank.

Who this role is not for

We'd rather say this up front than waste your time.

• If you want a ticket and a Figma file and to hand it back when it's done, this isn't the role.
• If AI is a chat window you paste code from, this isn't the role.
• If most of your career has been short client projects you never saw again after delivery, this probably isn't the role.

What you'll do

• Own features end to end. You take a founder's rough brief and turn it into a shipped feature. You ask the questions, write the plan, build it with your agents and watch it in production.
• Plan before you build. Most of your time goes into decisions: the data model, the failure cases, what to build and what to skip. The agent writes most of the code. You review every line that matters.
• Work on messy, real systems. Daily jobs that query several AI models, rate-limited APIs, scraping, queues, billing and our own MCP server. When something breaks, you find out why and make sure it doesn't happen again.
• Figure out the unfamiliar. A lot of what we build hasn't been built before. You research it, test the options and come back with a plan, not a question.

What you'll work on

• AI visibility tracking. Every day we run each customer's buyer questions through seven AI models and record which brands they name and which pages they cite.
• Citation Outreach tools. The tools our outreach team uses to find cited pages, reach publishers and track every placement, including an agent that drafts replies to publishers.
• Onsite content. An agent we're building that researches a keyword and writes an article ready to publish on the customer's own site.
• The stack. TypeScript, Node and Express, MongoDB, Redis and BullMQ on the back end. Next.js, React and Tailwind on the front end. The Claude and OpenAI APIs, Stripe and our own MCP server.

You might be who we need

If most of these sound like you, apply.

• You've worked as a software engineer for 7 years or more, and the back end is where you're strongest: APIs, databases, queues and the infrastructure under them.
• You design data models that hold up at scale, and you know how to find and fix a slow query.
• You've owned one product in production for a year or more: fixed it, scaled it and lived with your own decisions. Ideally at a product or SaaS company, or as an early engineer at a startup.
• You build with Claude Code, Cursor or Codex every day, and you've set them up properly: rules files, commands, MCP servers, subagents.
• You plan first. You can tell when the agent is confidently wrong, and you stop it before it ships.
• Give you something you've never built before and you research it, then come back with a plan.
• You write clear English: specs, pull requests and messages the founders can act on without a call.
• You live in Pakistan and want to work remotely with a team across Asia, Europe and the US.

MERN experience helps, since that's our stack. How you plan, solve problems and work with AI matters more.

What we expect from you

• Ownership. What you ship is yours. You check it in production, read the logs and fix what breaks without waiting to be asked.
• Speed. Be fast. Ship small, check it and ship the next thing. With agents doing the typing, a week-long task often isn't one.
• Learn as we go. Nobody has a playbook for AI search yet. We try things, measure them and change course quickly, so you'll learn something new every week.

What we offer

• Remote, anywhere in Pakistan. Any city is fine.
• Work with the founders. They build with Claude Code every day too. Most decisions take one conversation.
• Real ownership. Senior here means you decide how things get built, not just how fast. You'll shape how the engineering team works as it grows.
• Salary: $2,500–$3,000 a month.

No cover letter, and we tell you after every step whether you're moving forward.