AI Coding Task Engineer - Remote, $1,000 per Passing Task

Higlobe, Inc.; United States; Remote; USD 41,000 - 50,000; Full time

Benefits offered by this job

Remote work

Job summary

Higlobe, Inc. seeks experienced software engineers to create AI coding tasks from repos they know well. Each task pays US$1,000 upon passing review, and you can take on as many tasks as you like.

Remote work is available from any location, with a minimum of one task per week (around 10 hours). You’ll propose a change, build the task, write the brief, the correct solution and the grading checks, then stress-test and run model evaluations until the task is solid.

Qualifications

• 2+ years of professional software engineering.
• Code merged into a public open-source repository is preferred (or personal/private repo you own).
• Strong in at least one of Python, TypeScript, Go, Rust, Java or C++.
• Solid with Docker, Linux, Git and automated testing.
• Regular use of an AI coding tool.
• Experience writing tests, rubrics or evaluation criteria for AI models.
• Ability to explain a change clearly and concisely.
• At least 10 hours a week, starting now.
• A secure personal or work computer (no public/shared machines).

Responsibilities

• Turn a repository change into a task that AI models struggle with but is solvable by an engineer.
• Write the brief, the correct solution and the grading checks.
• Stress-test against lazy, incorrect and shortcut solutions.
• Run the task against leading AI models for evaluation.
• Submit revisions after reviewer feedback and repeat when needed.
• Produce subsequent tasks after acceptance.

Skills

Python

TypeScript

Go

Rust

Java

C++

Docker

Linux

Git

Automated testing

Tools

Docker

Linux

Git

Job description

Higlobe, Inc. seeks experienced software engineers to create AI coding tasks from repos they know well. Each task pays US$1,000 upon passing review, and you can take on as many tasks as you like.

Remote work is available from any location, with a minimum of one task per week (around 10 hours). You’ll propose a change, build the task, write the brief, the correct solution and the grading checks, then stress-test and run model evaluations until the task is solid.