iMerit launched its Scholars platform on 26 October 2026. It is a system for recruiting, onboarding, assessing, assigning, paying and keeping domain experts who evaluate, fine-tune and validate AI models. It is the company's first major launch since EXL completed its purchase of iMerit in July.
“We call it tormenting models”
Radha Basu, executive vice president and head of iMerit at EXL, to SiliconANGLE · siliconangle.com ↗
The Scholars community already counts more than 30,000 vetted professionals across 118 countries and 106 languages, with backgrounds in medicine, mathematics, agriculture, linguistics and autonomous systems. iMerit expects it to grow into the hundreds of thousands. The platform runs on its Ango Hub tooling for annotation, workflow automation and fine-tuning, handles multimodal data, and sends payments and compliance through Trolley's global payouts service. New features include matching experts by specialty and cognitive skill and secure workflows for sensitive data.
Experts pass three layers of vetting: a general AI and cognitive screen, a domain qualification that also tests how well they work with AI, and peer review by other experts. iMerit runs the projects itself, including quality checks and audits, instead of simply matching clients with freelancers. Clients can pay hourly by expertise, by deliverable or by outcome, and one customer once asked for about 400 extra workers within three or four days.
Portrait of Radha Basu, founder of iMerit, from TechCrunch's 2025 profile of the company
Image: TechCrunch
Radha Basu, who heads iMerit at EXL, says the industry is moving from building ever bigger foundation models to adapting them for specific, regulated uses in fields like healthcare, banking and insurance. Long-term relationships with experts matter, she says, because they build up knowledge of how models behave from one training cycle to the next. The company's biggest need is new areas of expertise, more languages and deeper levels of the expertise it has, and it plans to widen its coverage accordingly.
About iMerit
iMerit is a data annotation and generative-AI services company, founded in 2012, that supplies expert-led model evaluation and training data and builds its work on its Ango Hub tooling. It is now part of the services firm EXL.
It took its Scholars programme out of beta in July 2025, when it had more than 4,000 Scholars, presenting itself as an expert-led alternative to high-volume data labelling. On 24 June 2026 EXL announced a deal to acquire iMerit for up to $310 million, made up of $170 million upfront and $140 million in earnouts over two years. EXL said iMerit's Ango platform would be integrated into its agentic platforms.
2012
Founded
The work includes red-teaming, reinforcement learning from human feedback and chain-of-thought tracing. EXL agreed in June to buy iMerit for up to $310 million.