Author: Rob Laing, VP of Global Specialist Workforce
Years running high-growth tech companies have taught me about many things, from the nuances of language translation to the art of growing perfect flowering begonias in the controlled environment of an urban vertical farm. But across the board, in every role, one lesson arises again and again: technology can only be as good as the people behind it.
Now, as VP of Global Specialist Workforce at iMerit—a leader in data annotation and generative AI services—everything I’ve learned about the interaction of technology and humanity is finding new applications. And expansive new landscapes.
AI is opening more opportunities in the world than ever before. That much you already know. But the plot twist is this: AI has accelerated and expanded to the point that now it needs more help than ever from humans to continue its evolution.
The easy stuff has been done. The data’s been scooped. Now, it’s more crucial to have highly skilled humans guiding and training AI so it can move far beyond the realm of the rote and enter truly groundbreaking applications. iMerit has made a name for itself by training AI to help radiologists identify potential flags in patients’ imagery, help farmers detect invasive weeds via autonomous sensors, or allow self-driving vehicles to discern the difference between an inbound pedestrian and a passing reflection in a building window. Now, we’re training Large Language Models (LLMs) to clearly explain the chain of thought behind solving complex math puzzles, to make strategic recommendations that respect cultural nuance, or to use CoT reasoning to mimic the thought process of experienced physicians analyzing patient cases at scale.
For humans to give AI this level of advanced training, they need their special blend of critical thinking and specialized domain knowledge—not only within various fields and industries but also across languages, cultures, and geographic areas. Deep domain knowledge, explanatory precision, and rigorous logic are now required, with a careful, hands-on team management approach to match. Every nuance and detail of model training will have an amplified impact on the final scaled-up product, so diversity of expertise and precision in training is crucial. (We’ve all seen how “small” unintended biases within the data used to train AI can have terrible consequences in the long run. It can lead to an AI cancer detection application failing to detect cancers in darker-skinned patients after only being trained on lighter skin-tone images, or an autonomous vehicle software not knowing how to navigate rainy roads after being trained in drier climates, or an AI writing model assuming the title “CEO” would imply a male character.)
Typical crowdsourced training models simply aren’t equipped to deliver the accuracy and data quality iMerit’s clients need, nor do these models foster the caliber of workforce experience iMerit prides itself on. This is exactly why we developed our Scholars program—answering our clients’ needs as well as workers’.
Enter iMerit Scholars: Developing the Best Possible Product by Putting People First
For over ten years, iMerit has been a people-centric technology company. This strategy has catalyzed an eleven-year growth trajectory—with a diverse, balanced group of leaders at the helm and a workforce nurtured by a culture of mentorship, skilling, inclusion, and equitability.
When it comes to our full-time workforce, parental leave, scheduling flexibility, and advancement opportunities are more than benefits—they’re baked into the company’s way of working. They go hand in hand with sky-high retention rates of 90% as well as female workers (and female leadership) exceeding 50%. Few companies manage to remain truly multicultural as they scale, yet iMerit knows that the diverse thinking and ideas that multiculturalism yields are well worth the effort to uphold it.
iMerit Scholars builds on this same people-first ethos while bringing together a global team of expert consultants to help our clients evolve AI beyond basic data labeling into deeper, more specialized frontiers.
Scholars assemble custom teams of experts from around the world who are matched to specific AI development projects’ training requirements. Vetted and hand-selected by humans who understand humans, these experts can collaborate from anywhere and precisely, and reliably inform sophisticated AI development using their deep domain knowledge.
We can achieve this on a global scale that dynamically meets our client’s needs—while giving these workers a great experience. While the work we do is at technology’s bleeding edge, a quality work experience comes from a traditional human touch, experienced project management, and an insistence on clear, empathic, culturally aware communication. iMerit’s talent acquisition team and leadership team shine in each one of these areas. And, backed by worker-centric values, a history of ethical data services, a knowledge-driven culture, and our position as an early leader in AI, we know how to make the extraordinary happen. So, whether a project calls for a medical transcriptionist in France or a dozen PhDs in Mathematics who live in a specific region of India and who have experience in a nuanced sub-specialty, iMerit knows how to find them.
This solution starkly contrasts with a conventional crowdsourcing approach to training AI, which can not only lead to inconsistent quality and lack of expertise but also misses a key element—community. Scholars not only source qualified people, but also foster collaboration, offer mentorship, incorporate project management, and facilitate the sharing of institutional knowledge.
This model is an embodiment and evolution of everything I learned about “what works” distributing workflows across an online platform when I led a tech company that transformed language translation services. We grew a successful platform to be sure, but most valuable to me were the lessons I learned along the way—the ins, outs, pros, and cons of building and activating dynamic teams from around the globe. I learned how much it matters to manage teams well. And how the humans behind the usernames are the most important thing. I learned how much workers’ personal expertise and professional chops matter. How their lived experience and their livelihood correlated with the quality of our final product. How there was value in retention and relationships.
So when iMerit offered me the chance to help shape its new Scholars program, I knew all these learnings would be of service as we assemble and support teams of experts who understand the nuances of their professional domains. All the ingredients were there: brilliant people, a company that leads in AI and elevates its workforce, and the opportunity to shape the future of AI for the better.
These Scholars are more than contractors or consultants—they’re experts who bring invaluable training and lived experience to the table. And, in return, they deserve equitable compensation (that often far exceeds local norms), mentorship, hands-on management, and the chance to continually evolve their skills. These workers may be remote, but their experience with us is far from transactional. Clients often say that iMerit’s workforce acts as an extension of their team. This is why we’re building long-term relationships at Scholars because shortcuts don’t serve us.
Treat People Right and You’ll Get Your Product Right
For the past nine years, I ran an urban farm in New York City. We grew specialty produce in a local food industry where overwork, low pay, and a scant social safety net were the norm. But we knew treating our team well, including salaries and benefits, was not only the right thing to do—it was the key to retention and, ultimately, to a higher quality business and product. Long-term employee retention is rare in the hospitality world, and yet nine years in, even Amere, our bicycle delivery lead had stuck with us, providing for himself and his family while contributing to our culture and our success. His encyclopedic knowledge of chefs, products, and routes was immensely valuable, and something that only comes with time.
The seeds we plant in AI work the same way. What goes in comes out—the practices and teams used to train the technology will always be present in the resulting AI. So when our workforce—including 5,500 full-time employees plus our ever-growing cohort of specialized Scholars—has a liberating work experience, clients get great results and end-users experience a better final product as well.
Thus every hiring decision, whether for a scholar’s role on a specific project or a full-time role across projects, is an impactful one—on the quality of the AI that we train, and on people’s lives. Personally, this is what I may be most excited about: the second-order and third-order effects of thoughtful workforce recruitment and management.
Because ultimately, that’s what makes all this worth it. Pursuing knowledge, tapping into expertise, enriching the human experience, and achieving the remarkable. All it takes is prioritizing people to create the best AI imaginable.
