Trusted by
Meridial is an expert contractor marketplace powered by Invisible Technologies. We connect domain experts, in law, STEM, finance, coding, linguistics, safety, and other fields, with AI training projects that require specialized human knowledge and judgment.
Meridial is for subject matter experts across domains including law, medicine, coding, finance, creative writing, and others, who want to contribute and provide their services to frontier AI development.
Meridial experts provide critical feedback through traditional Reinforcement Learning from Human Feedback (RLHF) and Supervised Fine-tuning (SFT) processes like writing and answering prompts, evaluations, ranking responses, improving chain of thought reasoning, and red-teaming. Additionally, experts play a critical role in creating Reinforcement Learning (RL) environments, performing tasks like environment design, trajectory validation, judge model calibration, reward signal design, and edge case identification.
Experts create a profile, complete a brief confirmation step, and browse available opportunities for engagement. Those matched to a project sign a Statement of Work and contribute on their own schedule. There are currently 700+ engagements across 12 domains.
Earnings are set per project and determined by experience and expertise. Meridial publishes the rate range on each engagement, so experts can assess the opportunity before engaging.
Meridial sources engagements directly from the labs building frontier AI, rather than operating as an open marketplace. Domain expertise is matched with specific projects, giving experts access to cutting-edge work from top frontier models and enterprises.
