Full-Stack Engineer - AI Systems

Mercor; Mumbai; On-site; INR 1,200,000 - 1,800,000; Full time

Job summary

Mercor seeks experienced full-stack engineers to advance AI frontiers. You will write precise specifications, build complete reference implementations, and craft browser-driven acceptance criteria that verify real user-visible behavior.

You will calibrate task difficulty using frontier models and ensure features fail cleanly against an empty repository. The role requires 3+ years of full‑stack experience, fluency in frontend and backend technologies, and production-ready work on auth,

Qualifications

• 3+ years of professional full-stack engineering.
• Fluency in at least one frontend framework and one backend runtime.
• Experience implementing auth, permissions, payments, and file uploads in production.
• Ability to specify a system precisely in writing; criteria are the product.
• Comfort with Docker and git at commit level; CLI tooling.
• Experience building AI evaluation benchmarks or test suites is a plus.

Responsibilities

• Write product specifications precise enough for building the application and for automatic grading.
• Build a complete reference implementation that passes all criteria written.
• Author browser-driven acceptance criteria that verify real user-visible behavior.
• Calibrate task difficulty by running frontier models against the task and tuning.
• Verify tasks fail cleanly against an empty repository to ensure real feature complete.

Skills

Full-stack development

Frontend frameworks

Backend runtimes

Auth and permissions

Payment flows

File upload

Tools

Docker

Git

Job description

1. Role Overview

Mercor is seeking experienced full-stack engineers to work on an exciting opportunity to advance the frontier of AI.

2. Key Responsibilities

• Write product specifications precise enough that a competent engineer could build the application, and unambiguous enough to grade automatically

• Build a complete reference implementation of each specification that passes every criterion you have written

• Author browser-driven acceptance criteria that verify real user-visible behavior, not implementation details

• Calibrate task difficulty by running frontier models against your task and tuning until it lands in the target band

• Verify each task fails cleanly against an empty repository, so no criterion can be satisfied without genuinely building the feature

3. Ideal Qualifications

• 3+ years of professional full-stack engineering, having personally shipped and operated web applications

• Fluency in at least one modern frontend framework and one backend runtime, with working comfort across several

• Hands-on experience implementing auth, permissions, payment flows, and file upload in production, and knowing where each one usually breaks

• Ability to specify a system precisely in writing: the specification and its criteria are the product

• Comfort with Docker, git at commit level, and command-line tooling

• Prior experience building evaluations, benchmarks, or test suites for AI systems is a strong plus