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. In this spotlight, we spoke with Pradeep Das, Senior Staff Machine Learning Engineer at Vijil, whose career has focused on building AI systems at scale. Before joining Vijil, Pradeep worked on voice assistants at Viv Labs (Samsung) and developed large-scale recommendation systems and ML infrastructure at Amazon. Read on to learn more about Pradeep’s journey and how he’s helping Vijil make AI systems more trustworthy, reliable, and production-ready:

What you do at Vijil

At Vijil, we build trustworthy AI systems that  not only perform tasks reliably but also maintain alignment with human values under diverse conditions. I design multi-agent systems composed of specialized AI agents that cooperate to achieve their client’s goals: some generate decisions or actions, others verify their correctness, and others ensure the system’s compliance with organizational policies.

What inspired you to join Vijil, and how did your prior AI experience shape its vision?

I've always been drawn to building and scaling innovative technology. At Viv Labs (Samsung) previously, I worked on voice assistants, and saw the critical need for trust in AI applications. Then at Amazon, I worked on recommendation engines and ML infrastructure at a massive scale—another arena where reliability is paramount. When I learned about Vijil's mission to create reliable, secure, and safe AI agents, it felt like the perfect place to apply both my entrepreneurial and technical experiences.

My journey to Vijil came through a natural progression in my career working with increasingly sophisticated AI systems. Throughout my time at companies like Amazon and Samsung, I witnessed first-hand how AI was transforming from relatively simple rule-based systems to more complex learning models. To deepen my expertise in this rapidly evolving field, I completed several advanced NLP and ML courses at Stanford, focusing on grounding techniques and factuality in language models—foundational skills that would later prove crucial for building trustworthy AI systems.

What became apparent to me was that as these systems grew more capable, the challenges around reliability, security, and alignment became exponentially more important. At Amazon, I worked on scaling recommendation systems and building vector embedding infrastructure that could handle massive catalog expansions. This project required not just technical expertise but also strategic thinking about how to maintain performance while dramatically increasing the scale and complexity of our recommendation systems.

The emergence of generative AI in recent years represented a step change in capabilities, and with it came a whole new set of challenges around trustworthiness. When I learned about Vijil's mission to build the platform and tools to make trustworthy AI systems, it resonated deeply with the direction I saw the industry heading. The opportunity to help shape how organizations deploy AI responsibly felt like the right challenge at the right time.

What's compelling about Vijil is the opportunity to solve this trust problem not just for one company, but to deliver solutions that can benefit the entire industry and potentially establish a de facto standard. As developers integrate AI capabilities into mission-critical applications, guidelines and guardrails aren't optional– they become foundational requirements.

What’s one lesson you’ve learned from working in AI that you’ve applied directly to building Vijil?

Scale changes everything. Something I've seen repeatedly in my career is how solutions that work perfectly well in controlled environments often break down when deployed at massive scale or in diverse real-world settings.

When working on large technical systems, I've learned that scaling isn't just about adding more computing resources – it fundamentally changes how systems behave and the challenges they face. My background in distributed systems gives me a perspective that many in the ML field don't have. I understand not just the modeling aspects, but the entire system architecture required to make AI reliable at scale – from data processing pipelines to serving infrastructure to monitoring frameworks.

This distributed systems mindset is particularly relevant for trust in AI. It's trivial to build guardrails that perform well in laboratory conditions with carefully curated inputs. True robustness requires systems that can withstand the incredible diversity of real-world interactions, deliberate adversarial probing, and the countless unexpected edge cases that only emerge at scale. The principles that govern reliable distributed systems – redundancy, fault isolation, graceful degradation, and comprehensive monitoring – apply equally well to building trustworthy AI.

At Vijil, I'm applying these principles to build multi-agent systems that adhere to policies reliably. By designing architectures where specialized agents verify outputs, ground responses in established facts, and ensure proper entailment relationships, we're creating AI systems that remain trustworthy even as they scale to meet enterprise demands. This integration of distributed systems thinking with cutting-edge ML is what makes our approach particularly powerful for addressing real-world trust challenges.

What excites you about being part of the Vijil team?

I'm excited about the opportunity to define what trustworthy AI looks like in practice.

At Vijil, we’re building comprehensive frameworks for evaluating, improving, and monitoring AI trustworthiness. We’re exploring ways for models to reason through complex scenarios, not just pattern-match, and to verify their own outputs for factual accuracy and consistency.

It’s deeply satisfying to work on technology that will help shape how AI is deployed safely and responsibly across industries.

What are your personal goals for contributing to Vijil’s growth and success?

My goal is to help organizations see that security and reliability aren’t add-ons—they are foundational components that enhance AI’s value.

By developing multi-agent verification architectures into production-ready systems, I want Vijil to become the standard for responsible AI deployment. I’m focused on making trustworthy AI not just possible, but practical and scalable across industries.

How do you envision Vijil evolving over the next 5 years, and what role do you see yourself playing in that journey?

I see Vijil embedded in every major AI deployment, providing continuous monitoring, evaluation, and protection—just like cybersecurity does for software today.

As AI systems grow more powerful, Vijil will be at the forefront of making sure they remain robust, aligned, and trustworthy.

What do you see as the biggest opportunity for Vijil, and what challenges must be overcome to seize it?

Our biggest opportunity is becoming the essential trust layer for AI deployment across industries.

The main challenges are technical—building verification systems that work reliably across diverse models—and organizational: ensuring our trust mechanisms are easy to integrate into existing workflows without slowing teams down.

We also need to continue educating the market that trustworthy AI isn’t just about preventing harm—it’s about unlocking new value safely and responsibly.

What’s a widely held belief about AI that you question?

I question the belief that AI alignment can be solved solely through better algorithms.

Reliable systems emerge not from perfect components, but from robust architectures where specialized systems check and reinforce each other—just like in distributed systems or aircraft design.

The future of trustworthy AI will likely involve multi-agent architectures working in concert, rather than single, monolithic models trying to be perfect.

What is something about you that may surprise others?

I love challenging myself outdoors. I once completed the “Triple Tahoe”—three back-to-back marathons around Lake Tahoe—and have spent a lot of time hiking the Sierras and photographing the West Coast.

These experiences have taught me humility, patience, and a deeper appreciation for the complexity of both nature and technology.

👉 Want to learn more about Vijil and how we’re delivering mission-critical AI agents? Check out our website to learn more about the team, and set up a time to chat with us!