Prompt Engineer

Emergent is a platform for building full-stack web and mobile apps in minutes. It serves founders, builders, and teams who want to create and launch real applications quickly without starting from scratch. Users can sign in with Google, email, or SSO and use the product to turn ideas into working apps.

Emergent's agent is the product. Every capability it has, every edge case it handles, every interaction it gets right — traces back to the quality of the prompts behind it. This is a seat on the team that owns that quality directly.

You'll start by getting deep into the prompts already in production — understanding the standards, the patterns, and the reasoning behind how they're written. Once you've earned that context, you move to building new prompts for the agent. Day one is real work. The ramp is fast by design.

• The current prompt library — iterate on what's in production, understand the standards, and sharpen what's already there before you build anything new.
• New prompt development for the agent — once you've earned the context, you design and ship prompts that expand what the agent can do.
• Your own debugging loop — when something breaks or behaves unexpectedly, you own the investigation: codebase, scenarios, hypotheses, resolution.
• Prompt quality standards — as the team grows, your work sets the bar for how prompts are written and evaluated at Emergent.
• Speed and craft together — high scale means both matter; you hold both without trading one for the other.

• You've worked with LLMs or agents — even if briefly — and you've developed a feel for where models break and why.
• Curiosity is your default mode. When something doesn't work, you go looking before you go asking.
• You've operated in a fast-scaling startup where the pace was real and the stakes were high — Dunzo, Lovable, Replit, or something like them.
• You hit a wall in the docs and your first move is to dig into the codebase, map the possible scenarios, and form a view — not escalate.
• You have a strong instinct for how small changes in phrasing shift model behavior, and you find that interesting rather than tedious.
• You're from an IIT or a comparably rigorous technical background — or you can demonstrate the same depth of thinking without the credential.
• You've shipped things in environments where iteration speed mattered and quality couldn't slip.
• You're comfortable being the person who figures it out, not the person who waits to be told.

• Start with the problem, not the escalation. Dig into the codebase, map the scenarios, form a view — then bring it to the team if you're still stuck.
• Iterate before you invent. Understand what's already working and why before you build something new.
• Move at the pace the product demands. We deal with high scale; slow is not an option.
• Bring initiative. No one is going to hand you a task list. You see what needs doing and you do it.
• Collaborate with context. When you do bring something to the team, come with a point of view, not just a question.
• Stay curious. The models are changing fast. The best prompt engineers are the ones who stay ahead of what's possible.

₹20-25L cash plus ₹5L equity per year. On-site in Bengaluru with a team that built and scaled Dunzo, now building the agent that lets anyone on Earth create. Series C, high scale, fast pace — the work is real from day one.