As a human data solutions company, we help organizations build AI that is smarter, more reliable, and scalable, guided by humanity through multilingual data, evaluation, safety, and human oversight. DATAtalent: our network of 300k+ AI Contributors work tirelessly to ensure the data fueling AI frontier labs is of the highest quality, is culturally aware, and fit for purpose.
Adwaith Antony has four years of experience as a DATAtalent AI contributor, working across data annotation, evaluation, and model refinement projects with AI teams around the world. We asked him to reflect on his experience working as an AI Contributor – take a look into his insights below:
How did you first get involved in AI data work?
I started as a Software Engineer when AI was really gaining momentum. What stood out to me early on was how much time engineers were spending on data annotation and refinement. It made the importance of quality data very clear, very fast. I started volunteering for data collection tasks, saw the demand growing, and began picking up projects with vendors globally. The rest naturally followed from there.
What does a typical day look like?
Mostly remote and flexible. I start by checking emails, Slack, and LinkedIn for updates, then block out focused work sessions depending on what projects I’m working on at the time. Quality is everything in this work, so I’m deliberate about taking breaks (sometimes even a post-lunch nap!) to stay sharp. Some projects run on fixed schedules that fall at night in my time zone, and I use those evenings to upskill too: webinars, new tools, community spaces.
What’s been the most rewarding project?
Evaluating AI-generated videos. The task involved assessing cultural accuracy, language match, visual authenticity, things like whether the architecture, clothing, and landscape actually fit the target region. I have a personal interest in culture and languages, so it was a rare case where my own perspective felt genuinely valuable to the output. That sense of contribution being tangible made it stand out.
What does quality mean to you in this work?
Reliability. If someone reviews my output, it should be clear, well-reasoned, and aligned with the intent of the task, not just technically correct. I pay close attention to instructions at the start, track updates throughout, and know when to slow down. Some tasks need more thought than others, especially where context is nuanced. It’s about being dependable, not just done.
AI is advancing rapidly. How do you see the role evolving?
Toward higher-level judgment, not volume. As automation handles more routine annotation, what becomes valuable is the ability to evaluate and guide model behavior in complex, ambiguous situations, edge cases, unintended outputs, patterns of failure. The role will be less about isolated tasks and more about shaping better signals and improving feedback loops.
What qualities make a great AI contributor?
Attention to detail. Good judgment under ambiguity. Curiosity and a bit of healthy skepticism. But above all, the distinctly human things: reading intent, picking up emotional tone, understanding nuance. Those are what automation can’t replicate, and they’re what make the difference in the data.
Interested in DATAtalent?
Adwaith contributes through DATAtalent, DATAmundi’s AI contributor network. If you want to be involved in shaping the future of AI, join our network of expert AI contributors.
