Experienced Software Engineers: Coding Tasks for AI Systems

Terac; India; Remote; INR 8,009,000 - 13,349,000; Full time

Job summary

Terac is seeking experienced software engineers to design complex, self-contained problems for AI evaluation. You will configure development environments, work with Docker containers, and develop deterministic automated tests.

You will also craft reference solutions to ensure accurate grading. Ideal candidates thrive on independent work, navigate unfamiliar codebases, and demonstrate strong command-line skills and GitHub workflows.

Qualifications

• Proven experience as a software engineer.
• Strong command-line skills and GitHub workflows.
• Ability to navigate unfamiliar codebases independently and write tests.

Responsibilities

• Design realistic, self-contained software problems for AI evaluation.
• Configure development environments and Docker containers.
• Write robust, deterministic automated tests.
• Create verified reference solutions for reliable grading.

Skills

Command-line tools

GitHub workflows

Codebase navigation

Tools

Docker

Git

Job description

What We're Researching

We are running a paid project focused on creating and validating complex coding tasks for AI systems. The goal is to build realistic, self-contained software engineering problems that push the boundaries of automated coding agents. Your contributions will directly improve how these systems navigate real-world development environments.

How It Works

You will design intricate software problems that an AI agent must solve within a remote terminal environment. This involves implementing features, debugging codebases, configuring Docker containers, and writing deterministic automated tests. You will also create reference solutions to ensure every task can be graded reliably and accurately. Participants work independently to verify that each problem is challenging, solvable, and completely self-contained.

Who This Is For

We welcome experienced software engineers who are highly comfortable working independently in unfamiliar codebases. You should have strong command-line skills, proficiency with GitHub, and experience writing clean tests. Candidates with a background in complex debugging, system configuration, or automated testing are highly encouraged to apply.

What You'll Do

• Design realistic, self-contained software engineering problems for AI evaluation.
• Configure development environments and work within Docker containers.
• Write robust, deterministic automated tests to evaluate task completion.
• Create verified reference solutions to ensure tasks are accurately graded.

Who Should Apply

• Professional experience working as a software engineer.
• Strong proficiency with command-line tools and GitHub workflows.
• Comfortable navigating and debugging unfamiliar codebases independently.
• Willingness to complete an identity verification and skills assessment.

Compensation

$82 per task

About Terac

Terac is building the world's largest pool of vetted human experts for AI. Researchers, AI labs, and product teams use Terac to recruit, screen, and pay study participants across industries, languages, and skill sets.

Learn more at or on YouTube at @jointerac.