1. Your Story
Introduce yourself and your role. What problem were you compelled to solve?
I’ve spent two decades as an AI product leader at companies like Snowflake, Google, LiveRamp, and eBay — leading teams of up to 450 people and creating products that generated over $2 billion in incremental revenue. The problem that has driven my entire career is the gap between what AI can technically do and what organizations can actually use. I’ve seen brilliant AI models sit on shelves because nobody thought about the human workflow. My work has always been about closing that gap.
I grew up in Sri Lanka, where I got hooked on technology by asking my grandmother for batteries and wires to build electrical inventions. That same curiosity led me from studying Electrical Engineering at Yale and Stanford to business at Harvard — and it still drives my AI work today.
2. How You Operate
What is your operating model — do you build in-house teams, work with external partners, or a hybrid approach?
I build cross-functional teams that bring together product, engineering, data science, and go-to-market. The scale has ranged from a 30-person startup where I was CPO to large enterprises. The principles don’t change with size — I call it “psychological safety with clear outcomes.” Trust people on the how, be rigorous about the what and why. The framework scales — the team size doesn’t define it.
3. Standing Apart
In a competitive market, what differentiates your approach?
Deep technical fluency combined with strategic thinking. I started my career programming robot vision systems, then spent 20 years at the intersection of AI, data, and enterprise software. The hardest problems in AI product management aren’t technical — they’re translation problems. I can sit with a data science team and understand model architecture, then walk into a boardroom and explain why it matters. Most people live on one side of that divide. I’ve built my career in the middle.
4. Who You Serve
Which industries or sectors do you primarily work in, and how has that focus evolved over time?
Enterprise software, marketing technology, e-commerce, data platforms, and analytics. The evolution tracks the AI stack itself — machine learning for ad optimization early on, deep learning for modeled conversions at Google, and large language models for natural language data access. Each layer got harder and more valuable. The industries change. The underlying challenge stays the same: helping organizations make better decisions with their data.
5. What You’re Known For
What do organizations typically come to you for?
Turning AI from a demo into a business. I build “zero-to-one” AI products that create new revenue streams, and I make AI semantically understand business context — not just process data, but comprehend what it means. I’ve developed a framework I call the 5Ps of Product — plan, problem, product, promotion, and platform — that shapes how I approach every product I build.
6. Staying Ahead
How do you personally stay ahead of industry shifts when insights and data age so quickly?
I stay close to the technology — I still prototype and build using AI-assisted coding, which connects me back to my engineering roots. But the real edge is cross-domain pattern recognition. Having worked across advertising, e-commerce, and data platforms — and having traveled to over 80 countries on five continents — I’ve learned that the same problems resurface in different industries wearing different clothes. That global perspective is something no desk research can replicate.
7. Experience & Satisfaction
How do you measure success beyond the deliverables?
Adoption rates. The best product in the world is worthless if nobody uses it. Beyond adoption, I measure durability — did the product keep growing after I moved on? Is the team stronger for having worked together? At Google, the core measurement products I built continue to increase in value after my tenure. That’s the kind of success that matters most.
8. Beyond the Role
What kind of involvement do you maintain beyond your primary work?
I’m passionate about access to education because scholarships fueled my own journey. I co-founded Inspire Inc., a nonprofit that has impacted over 300,000 students through volunteer consulting to education leaders, and I sit on the board of Breakthrough SF, supporting underserved youth on the path to college. I also share frameworks and thinking at onilgunawardana.com, and helped create Google’s Product Manager Circle mentoring program, which grew to 700+ participants.
9. Choosing the Right Fit
Have you ever turned down opportunities? What makes the right fit?
Yes. I need a genuine problem worth solving, leadership committed to doing the work, and organizational willingness to let AI change how people actually work — not just add a layer on top. The projects that fail aren’t the ones with insufficient budgets. They’re the ones where the organization isn’t ready to use what gets built.
10. Navigating Challenges
What has been one of the most defining challenges you’ve faced, and how did you overcome it?
At Snowflake, the challenge was making large language models accurate enough for enterprise data, complex, messy business data that breaks most AI tools. Ask about revenue by region, and the LLM needs to understand specific schemas, naming conventions, and business logic. Generic models fail spectacularly.
We built a semantic layer — Semantic Views — that gave the AI a structured understanding of what the data actually meant. It increased accuracy by 60%. The hardest AI problems aren’t model problems — they’re understanding problems. No amount of compute compensates if the system doesn’t understand the context.
11. Innovation & Evolution
How do you foster innovation within your teams while staying adaptable to emerging trends?
First-principles thinking. Strip problems back to fundamentals before jumping to solutions. Then hire curious people, give them problems worth solving rather than features to build, and protect their time from organizational noise. I also encourage what I call “adjacent exploration” — studying fields outside your domain. Some of our best product ideas came from people applying patterns they’d seen in completely different industries.
12. Culture as Strategy
What role does culture play in your success, and how do you build it?
Culture determines whether an AI product succeeds or fails. I build it around three principles: trust and openness, intellectual honesty, and clarity of purpose — everyone should articulate why their work matters in one sentence. At AT&T Interactive, this approach produced the best product launch in company history — a $70 million hyperlocal advertising product born from conversations with 250+ small businesses.
13. Vision Forward
Where do you see your field heading in the next 5–10 years? What goals are guiding you today?
The next decade of AI will be defined by trust. The most important AI products won’t be the most technically sophisticated — they’ll be the ones that best understand human context and workflow. That’s what I’ve spent my career building toward as an AI product leader, and it’s never been more relevant.
Talent is everywhere. Opportunity isn’t. The best technology should amplify everyone, not just those who already have advantages.