I build products that turn user frustration into business outcomes — across AI, consumer, and enterprise.
I'm a PM who works at the intersection of user behavior, data, and business strategy. Over the last 5 years, I've shipped agentic AI platforms for enterprise insurtech, designed freemium-to-paid conversion systems for large-scale consumer products, and rebuilt retention loops for healthtech and B2B SaaS companies.
The thing I'm probably best at: staying connected to the user problem while making the business case at the same time. I can go from a 15,000-review NLP analysis to a pricing model to a sprint plan — and not lose the thread in between.
These aren't values I list on a slide. They're decisions I make regularly, sometimes uncomfortably.
Fall in love with the problem, not the solution.
Before writing a user story, I try to articulate the problem in one sentence from the user's perspective and one from the business's. If those two don't connect, I go back to discovery. Every case study in this portfolio started this way — and in a few of them, the original brief changed significantly once I did.
Define success before you design anything.
"The feature launched" is not a success metric. "Weekly active adjusters who resolved a claim in under 5 minutes increased by 40%" is. I've seen too many teams ship things and then argue about whether they won. I'd rather have that argument before the sprint starts.
Clarity over consensus.
I'd rather make a clear, well-reasoned call that some stakeholders push back on than ship a vague compromise. Misalignment that gets papered over doesn't go away — it becomes technical debt and strategic confusion down the line.
AI is a capability, not a category.
I evaluate AI/ML opportunities through three questions: Does it reduce a friction that actually matters to users? Does it get better with data over time? Can we explain its behavior to the people who rely on it? If the answer to all three isn't yes, I'm skeptical of the use case regardless of how technically impressive it is.
Ship to learn, not to launch.
Every release is a hypothesis test. I keep retrospectives honest — including the ones where the metric didn't move and we have to figure out why.
| Domain experience | AI/ML products, D2C consumer, B2B SaaS, Insurtech, Healthtech, EdTech |
| Stage experience | 0→1 builds · growth-stage scaling · enterprise |
| Technical depth | Reads code · writes SQL · designs ML evaluation frameworks · defines agent architectures |
| Key collaborators | Data science · Design · Engineering · Legal/Compliance · Sales · Leadership |
| Research methods | NLP review mining · user interviews · funnel analytics · session recordings · A/B experimentation |