Human-first, AI-enabled: frontier models support the research, simulations, and publications behind the practice, while T provides the direction, judgment, and domain expertise.
I.Translation over prediction.
The job isn't to forecast. It's to translate what an insurance policy offers into capital and risk implications a capital allocator can act on.
For example —
- Collapse standalone insurance placements into master programs. Delivers scale for property insurance during construction and operation.
- Create a consistent methodology for sizing policy limits. Stays inside risk appetite without leaving capital on the table.
- Deliver security posting via surety bonds and surety-backed letters of credit. Reduces cost of capital.
II.Curious by habit.
The tools that help me translate keep changing. So I learn to build with whatever's new, see what it can carry, see what it can't, then revisit the question I was asking later.
In 2017: While learning about blockchain, I wrote a LinkedIn article imagining a conversation between me (in 2017) and a chatbot (2050 version of me). What felt radical in 2017 doesn't seem far-fetched given how far we have advanced with voice and image generation.
In 2020: I learned to build a Q&A chatbot as an experiment to earn the Microsoft Azure AI certification.
In 2022: ChatGPT 3.5 arrived and the steep learning curve began.
In 2023: Claude arrived.
The portfolio below is what's come out of the curiosity habit so far.