In our Team Spotlight series, we sit down with different members of the Vijil team to hear their stories, what drives them, and how they’re shaping the future of AI. For this spotlight, we spoke with Subho Majumdar, PhD, co-founder and Head of AI at Vijil. Before joining Vijil, Subho built responsible AI practices at AT&T, Twitch, Splunk, and Geico and wrote the book (literally) on trustworthy machine learning. Read on to learn more about his journey and how he’s helping Vijil build trust into AI agents.

What led you to Vijil?

My journey into trustworthy machine learning is that of a researcher turned practitioner. Back in 2019 at AT&T Labs, I saw a major gap between the research and the practice of trustworthy machine learning: while academic research on fairness and security was advancing, almost none of it was translating into real-world tools that practitioners could use easily.

At that time, businesses were hesitant to invest in responsible AI, believing that making models more fair or transparent would lower accuracy, and ultimately, business value. But my team at AT&T Labs proved that this wasn’t true. Fairness is inseparable from correctness.

Even so, responsible AI teams were often seen as cost centers rather than value drivers. That became painfully clear when my team at Twitch got cut in 2023.

Then, Generative AI changed everything. Suddenly, businesses saw AI as the next generation of SaaS. With AI agents becoming more capable, security, safety, and compliance were no longer optional—they became essential. Just like cybersecurity, AI security became a first-order business priority that blocks the deployment of AI agents into production.

That’s when my next step became clear: rather than working on responsible AI for just one company, I wanted to help as many companies as possible build AI systems that are reliable, secure, and production-ready. That’s why Vijil is such a compelling opportunity.

What’s one lesson you’re bringing from your experience?

Listen to customers. AI teams focus too much on the promise of the technology rather than its perils, and focus on their idea of a solution rather than the problems of the people they serve.

At Vijil, we don’t define “trust” in a vacuum. We construct it in a dialog with customers based on their priorities rather than some arcane academic standards. Every AI system has unique risks, and we need to tailor a solution to match its specific needs. Doing this at scale is the challenge.

What excites you most about Vijil?

The experience and expertise of this team. I get to work alongside some of the most accomplished business and technical leaders in AI, and I learn something new every day. By the same token, so to speak, I have the opportunity to be a mentor to some of the most talented young researchers and engineers I’ve worked with. I value the privilege of being able to guide their work and drive research on behalf of Vijil. I’m proud to have co-authored 3 papers in the past year that we presented at top-tier conferences. Many more to come.

Where do you see Vijil in the next five years?

Five years is a long time in AI, but two key priorities stand out:

First, broadening our impact by serving more customers—helping many more build and use trusted agents sooner. From day one, this has been a core focus. If we did nothing else, I’d consider it a success.

Second, establishing the basis of truly trustworthy AI agents. Right now, very few people are talking about the trustworthiness of AI agents. As AI agents evolve into autonomous agents, we cannot afford to ignore their trustworthiness. But trust is not a one-size-fits-all rack coat—it has to be a bespoke tailored suit that’s specific to the business context. What matters most for one business might not matter as much to another. Vijil is creating the tools to make it possible for every agent in every environment to be trusted in its role and task.

What’s one thing about you that people wouldn’t know from your LinkedIn profile?

I’m a huge foodie. I love to cook, and my wife runs her own Indian-inspired ice cream brand in Seattle.

Want to learn more about Vijil and how we’re making AI more secure, reliable, and production-ready? Check out our website to learn more about the team, and set up a time to chat with us!