If you’re the type who can make Bash sing and make broken systems confess, this is for you. You’ll stress-test advanced AI models with real software engineering scenarios, then document what fails so the model gets sharper, safer, and more reliable.
About Invisible
Invisible supports AI development by creating high-quality training data and expert evaluation workflows. They partner with technical specialists to validate outputs, identify failure modes, and improve model reasoning for real-world use.
Schedule
Contract, freelance, remote (U.S. listed).
Hours vary based on your availability and project demand.
You provide a secure computer and high-speed internet connection.
What You’ll Do
⦁ Converse with AI models on software engineering tasks and technical scenarios using Bash
⦁ Verify logical correctness, Bash fluency, and solution completeness
⦁ Assess code quality for clarity, maintainability, and secure practices
⦁ Capture reproducible error traces and document recurring failure patterns
⦁ Challenge the model on complex topics (APIs, debugging, distributed systems, systems programming) and flag weak reasoning
⦁ Suggest improvements to prompts, evaluation criteria, and quality metrics based on what you observe
What You Need
⦁ Strong Bash scripting skills with real-world engineering experience
⦁ Solid understanding of core CS fundamentals (algorithms, data structures, architecture concepts)
⦁ Comfort working across real engineering scenarios (debugging, security, system behavior, integrations)
⦁ Ability to clearly explain reasoning and “show your work” in writing
⦁ High attention to detail and consistency when evaluating technical outputs
⦁ Degree in CS/Software Engineering (BS/MS/PhD) is ideal; open-source work, technical writing, or equivalent experience strongly valued
Benefits
⦁ Flexible, remote contract work
⦁ Pay range of $8–$65/hour (rate based on experience, expertise, and location)
⦁ Hands-on work with cutting-edge AI tools and evaluation pipelines
⦁ Direct impact on model quality, safety, and usefulness
⦁ Project-based opportunities that can expand with performance and availability
Strong reviewers don’t sit in the queue long.
If you’re ready to put your Bash expertise to work and help train the models people will rely on, apply now and set a rate that matches your skill.
Happy Hunting,
~Two Chicks…