Software Engineer

JuiceLabs AI; Bengaluru; On-site; INR 1,200,000 - 1,800,000; Full time

Job summary

JuiceLabs AI is looking for a backend/system design engineer to work directly with the Cofounder/CTO on AI-centric products. You will define service boundaries, data models and architectural trade-offs before code, own the specs and design docs, and review outputs for correctness.

You’ll set up workflows with CLAUDE.md and agent tooling to ensure reliable delivery. We value high agency, deep system design, and a production mindset.

Qualifications

• High agency and ability to close gaps without direction.
• Real system design depth: reasoning about consistency, caching and scale.
• Clear technical writing for PRDs and design docs.
• Production experience: deployed systems, monitored, paged, fixed.
• Experience building AI-enabled tools and agent workflows.
• 2–6 years of full-time professional experience.

Responsibilities

• Design the system: service boundaries, data models, failure modes.
• Write the spec: PRDs and technical designs as the source of truth.
• Review PRs: human, agent, and model outputs for correctness.
• Set the workflow: integrate CLAUDE.md, AGENTS.md, subagents, and custom commands.
• Drive AI tooling strategies and maintain production-ready code.

Skills

High agency

Software architecture

System design

Technical writing

Python

PostgreSQL

Cloud infrastructure

Docker

CI/CD

Tools

FastAPI

Postgres

GCP

AWS

Docker

Job description

Where creative engineering meets applied AI.

We build vertical, AI-native tools that unlock fresh insights and creative superpowers for clients in advertising, ecommerce, and beyond. We're a nimble team of founders and builders obsessed with using AI to solve real problems in ways that didn't exist eighteen months ago.

The Opportunity

Six months ago, we would have posted this as a backend role. That role doesn't really exist any more.

Writing code has become the cheapest part of building software. Claude Code and Codex do the typing, and they do it fast. What they can't do is decide what's worth building, hold an entire system in their head, or notice that the very confident thing they just produced will quietly fall over at 10x traffic.

That layer - the one above the code - is the job.

You'll work directly with our Cofounder/CTO on products going from zero to one, with AI at the core of both what we build and how we build it. Real ownership, real scope, and an unusual amount of leverage: one engineer with good judgment and a well-tuned agent setup now ships what a small team used to.

What you'll do

Design the system. Service boundaries, data models, failure modes, what runs where and why. You'll make architectural calls before a single line gets generated, and you'll be the person who can explain the trade-off six months later.

Write the spec. PRDs and technical designs are the new source code. A vague spec now produces confidently wrong software at remarkable speed. We're looking for people who can think clearly on paper — scope, edge cases, non-goals, what "done" means.

Review everything. Human PRs, agent PRs, your own output at 2am. You're the last line of defence between "it runs" and "it's correct." This is where deep engineering knowledge actually pays off now — spotting the N+1, the race condition, the abstraction that will hurt in three months.

Set the workflow. Bring your harness. CLAUDE.md, AGENTS.md, subagents, custom commands, eval loops, whatever you've built to make agents reliable instead of merely impressive. We want to steal your good ideas.

What we're looking for

High agency. You see the gap and close it. You don't wait for direction, and you don't need to be chased for follow-through. This is the single trait we screen hardest for.

Real system design depth. You can reason about consistency, caching, queues, and scale - and about complexity, because reviewing generated code means knowing when an elegant-looking solution is quadratic. LeetCode ninja not required. Being able to explain why this won't blow up at 10x, very much required.

You write well. Clear PRDs, clear design docs, clear PR comments. Writing is thinking, and it's most of how you'll direct both humans and models. Comfortable reading PRDs with technical jargon.

You've shipped to production. Deployed things, monitored them, been paged by them, fixed them. Docker, CI/CD, cloud infra. Our stack is mostly Python (FastAPI), Postgres, GCP/AWS — but we care more about how you think than which framework you've memorised.

You actually drive AI tools. Not "I've tried Copilot." We mean you've developed opinions: where agents excel, where they confidently lie, how you structure a repo so they don't wander, how you verify what comes back.

Startup energy. Multiple hats, fast cycles, and the judgment to balance "ship it now" against "will this survive contact with real users?"

2-6 yrs professional experience (full-time positions, excluding internships)

Interview - what to expect:

• Show us what you've built - personal projects, internal tools you can show
• Show us your workflows - how do you build high-quality features and fast?
• Questions on architecture & system design