iMerit Launches Scholars Platform to Empower AI Developers
IMerit Technology Service Pvt Ltd., a provider of data annotation and artificial intelligence training services, which was acquired by ExlService Holdings Inc. in July, has launched a new platform aimed at assisting AI developers in recruiting and managing specialized experts for evaluating, fine-tuning, and validating models intended for demanding enterprise applications.
The iMerit Scholars platform is specifically designed to manage a global network that the company anticipates expanding to encompass hundreds of thousands of experts. Currently, its community includes over 30,000 vetted professionals spanning 118 countries and proficient in 106 languages across various disciplines, including medicine, mathematics, agriculture, and autonomous systems.
This launch signifies a strategic shift in AI development, moving away from solely building powerful foundational models to adapting these models for targeted applications, especially within regulated sectors such as healthcare, banking, and insurance.
“It’s like having a product and saying, ‘What are we going to do with it?’” explained Radha Basu, Executive Vice President and Head of iMerit at EXL. She emphasized that specialized human expertise is increasingly vital as enterprises integrate AI into applications where mistakes could lead to serious repercussions.
The platform offers centralized tools for recruiting, onboarding, evaluating, assigning, and compensating experts. Built upon iMerit’s proprietary Ango Hub data annotation, workflow automation, and model fine-tuning platform, it facilitates collaboration on multimodal training data and streamlines model evaluation and feedback processes. Payment processing and compliance features are provided through integration with Trolley Technologies Inc.
In contrast to traditional data labeling services, iMerit underscores the specialists’ role in challenging AI models and identifying errors that automated assessments might overlook. “We call it tormenting models,” Basu remarked. “Question the model, question the model to make sure it’s working, actually try to break it.”
The vetting process for participants involves a rigorous three-tier assessment. The initial screening evaluates general AI capabilities and cognitive skills, while subsequent assessments focus on domain-specific qualifications and the candidate’s ability to effectively collaborate with AI models. The final evaluations are conducted through peer reviews from other experts.
iMerit actively manages customer projects rather than merely connecting clients with independent contractors. The company’s responsibilities encompass project management, quality assurance, and auditing. Experts are compensated for their individual engagements but are also integral members of the broader Scholars community.
Establishing long-term relationships is crucial, according to Basu, as specialists accumulate knowledge about models through successive evaluation and training cycles. The company offers various pricing options, including hourly rates based on expertise, contracts linked to specific deliverables, and engagements tied to measurable outcomes. Basu cited agricultural models that identify early crop failures as an example where evaluations could be measured against performance targets.
iMerit also boasts the capacity to handle sudden surges in demand. Basu noted an instance where a client requested around 400 additional workers within three to four days to meet an urgent need. The EXL acquisition provides iMerit with enhanced access to healthcare and insurance expertise, further bolstering its capacity to support enterprise-specific AI applications.
As part of its ongoing strategy, the company is keen on expanding its coverage of specialized disciplines and languages. “Our biggest need right now is to go into new expertise, scale into newer languages, and into newer levels of the same expertise,” Basu concluded.