Leona Guo🇺🇸

About Leona Guo

Head of AI Adoption at AlphaX Re Capital

I started my career in 2019 at WePay, a payments company acquired by Chase, where I worked on small-business transaction systems. Over four years, I grew from a junior engineer to a staff engineer. I started with accounting reconciliation automation and eventually worked on large-scale, real-time transaction infrastructure connecting digital systems with physical payment terminals. The goal was simple but demanding: whether a customer swiped a card or completed an online transaction, it should appear in their account activity within a minute.

Those four years taught me how strong software engineering teams operate—how to build reliable systems, collaborate across functions, and grow quickly when given enough ownership.

I then joined a startup incubator focused on pre-seed companies in Silicon Valley. We reviewed roughly 500 startups each year, which gave me a front-row seat to both successful companies and failed experiments. I also helped manage the advisor community, connecting founders with experienced operators, investors, and industry experts. At the same time, we regularly hosted entrepreneur groups from Asia—successful business owners coming to Silicon Valley to understand AI and the next wave of technology.

After two years, I decided to build again.

As a founding engineer, I worked on what we described as an AI-native LinkedIn: a product that enriched your first- and second-degree network with public background information, understood what problem you were trying to solve through conversation, recommended the right people to talk to, and helped make the introduction. Because the recommendations came from trusted networks, the probability of a meaningful connection was much higher.

We built a working POC within four months and received investment from South Park Commons.

One of the hardest technical problems was search quality. At the time, the dominant approach was RAG, but RAG plus a single LLM was not good enough at understanding nuanced intent. Search results were often inaccurate or repeatedly surfaced the same people. Soon after Anthropic began publishing ideas around multi-agent systems, I experimented with a multi-agent search architecture. Within weeks, recommendation quality improved significantly. That experience shaped how I think about AI engineering: when an important problem is not solved by the standard stack, go back to the frontier and see whether new research changes what is possible.

Most recently, I took those experiences into the physical economy.

Over the past year, I have been deeply embedded in a real estate development and construction company, where I saw how much operational friction still exists in traditional businesses: fragmented workflows, siloed information, manual reporting, and processes that often live only in people's heads.

My work has increasingly become a combination of business transformation and engineering—helping management clarify workflows while using AI to turn ideas into working products, processes, and tools as quickly as possible. I have spent the year learning the industry from the inside while continuously refining the systems around it.

Throughout all of this, there has always been a small voice in the back of my head saying: look at how fast the world is changing—why are you still doing things the old way?

That voice keeps pushing me to change how I work, move faster, and use limited time to produce more interesting outcomes.

One thing I have learned over the past seven years is that the quality of a decision is not always obvious in the short term. Sometimes the best path is simply to keep taking small steps, learn from each one, and use that progress as energy for the next.

That is my story so far. I would love to compare notes, and I hope some part of it is useful to you.

Expertise

Sessions with Leona Guo

• General Mentorship Session (30 min)