Software Engineering AI Task Author

The Work

You will create and calibrate realistic software engineering challenges used to train and evaluate AI agents. Each task should reflect a plausible production problem, have the right level of difficulty, and include outcomes that can be checked.

You will also solve each task yourself, apply and verify the fix, and review AI-drafted prompts or scoring rubrics.

• Write task prompts based on production-style failures and debugging workflows.
• Build task environments with codebases, logs, alerts, and system states.
• Create rubrics covering the correct diagnosis, fix, and validation criteria.
• Develop tasks involving integration failures, log-based debugging, source-code bugs, and configuration changes.
• Adjust code complexity, log volume, failure modes, and ambiguity to calibrate difficulty.
• Implement and verify fixes in live or simulated environments.

What It Pays and Takes

This is a part-time contractor role for candidates based in Canada who can work in English. The listing marks it as entry level, but the required skills call for hands-on software engineering experience in industry or production environments.

• Pay: $70-$120 USD per hour.
• Time: 20+ hours per week.
• Location: Canada only.
• Language: English.
• Work type: Contractor and part time.
• Required: Experience debugging with logs, observability tools, and code tracing.
• Required: Fluency with Git, version control workflows, and code review practices.
• Required: Experience with infrastructure, integrations, APIs, services, configuration files, or CI/CD systems.
• Required: Ability to write and verify fixes and define precise, checkable engineering outcomes.
• Helpful: Experience in enterprise SaaS, e-commerce, telecommunications, or industrial engineering.
• Helpful: Familiarity with Datadog, PagerDuty, Grafana, or cloud-native systems on AWS, GCP, or Azure.

How It Works

OpenTrain AI is the hiring and contracting organization for this role. You apply on OpenTrain with your resume, then complete the application on the hiring site.

About AI Training Work

AI training is the human work behind systems that learn from examples, including writing tasks, rating model behavior, and checking whether AI-generated solutions work. OpenTrain helps people find and build careers in this field, where software engineering experience is used to create reliable training and evaluation data.