Safety AI Data Trainer
Safety AI Data Trainer
Join the AI revolution: work with the team that’s supporting the latest cutting-edge Large Language Models.
Invisible Technologies is dedicated to fusing human creativity and intuition with cutting-edge technology to create a future rich in impact and meaning. We firmly believe that human involvement is essential in unlocking the full potential of AI, ensuring it is developed with greater accuracy, quality, safety, reliability, and fairness. This human-centric approach underpins our unique process orchestration engine, a platform that merges artificial and human intelligence with automation to remove operational bottlenecks and pave the way for growth and innovation in our clients.
The Business Context
You already use AI in many ways—like deciding what products and services to order—and it may be most familiar to you as a chatbot, as an avatar-maker, or as a way to unlock your screen. But here’s what AI may be able to help the world with: finding medical diagnoses, teaching you about scientific research, and calculating the complexities of any function.
But, like humans, AI models are what they eat. They’re only as good as the rules they know and the data they’re trained on. We’re the team that helps shape these behaviors.
This work specifically involves exposure to varying kinds of sensitive and explicit content, via text and possibly image. This may include, but not be limited to, written depictions of sexual content, violence, and medical content. There will also be work involving representational bias.
Safety AI Data Training work is for the purposes of improving model alignment against safety standards and user policies. Our goal is to protect a wider audience from harm while creating a supportive and sustainable environment for those engaging in this important work.
These tasks call for a great deal of cognitive flexibility and creativity. These tasks will require both a) generating data through complex prompt engineering and b) reviewing and annotating that data.
- Access to our mental health support platform providing therapy, emergency counseling, and crisis support
- Regular check-ins
- Wellness surveys
- Opt out with the possibility of paid alternatives
Task availability supports work for 20 hours per week.
Who We’re Looking For
This is a technology job that relies more fully on your competencies than your formal experience. You’re the sort of person who is exceptional at both generating and editing copy. You’ve also got the metacognitive awareness to “show your work.” In addition, you’ve cultivated a deep interest in linguistics and unwavering curiosity. You know how to both read between the lines and pattern-spot. You instinctively document learning and continuously refine methods. Ideally, you’ve spent a great deal of time already playing with language models, and you consider yourself to be a casual prompt engineer.
You are a(n):
- Individual that is comfortable being exposed to sensitive or explicit content
- Exceptional writer and a skilled grammarian
- Diligent Reader
- Perpetual Learner
- Instinctive Educator
- Exceptional problem-solver, able to self-teach
- Motivated individual with mental agility to stay in context in an ever-changing environment
- Deeply curious about how AI models work and how they can be improved
- Committed to self-care to manage the emotional demands of the role.
- Exceptional written and spoken English
- Resident of US or Mexico
Nice to have:
- Creative writing education, experience, or interest
- Previous work with sensitive content
- Research and data analysis
- Editorial experience
- Experience working with Gsuite and proficient in Excel/Sheets
- Prompt engineering experience
Compensation will be $22.75-26.00 an hour.