Vivek Pawale
The person who has been building this, in one form or another, since before it was a company.
In 1999, Vivek Pawale finished a PhD in math at Brandeis and took a job that didn't have a name yet. The company was called AIR — Applied Insurance Research — and what they needed was someone who could stand between the scientists who understood hurricanes and the engineers who needed to turn that understanding into software. His own description of the role, years later: "I'm neither a meteorologist, nor a seismologist, nor a programmer. So what am I?" The answer turned out to be: whatever the industry needed next, for the next twenty-five years.
At AIR, Vivek spent five years elbow-deep in raw catastrophe data — hurricanes, earthquakes, whatever came in — building the statistical curves that became the industry's hurricane models, back when there was no Python and MATLAB was barely out of diapers. He wrote the models himself, in Fortran, including the spline function he still brings up when someone suggests a modern toolkit could have done it faster.
He's only half joking.
He left AIR in 2004 and landed on Munich Re's weather trading desk in New York, pricing weather derivative contracts alongside a trader named Paul. Then Katrina hit, the desk's book took the same losses every book in the market took that year, and Munich Re decided to get out of the business. Marty — who was rebuilding his own desk after his previous trader left — called Vivek in 2006. He was Marty's third hire. They've worked together every year since.
What Vivek took from that year wasn't the loss. It was a rule: it's not what your model does that matters, it's what your model doesn't do. You have to know the edges of your own model and build around them — through policy terms, through limits, through the payout structure itself. Twenty years later, it's still the first thing he'll tell you about how ClimateDelta is built.
The years since have taken him through Endurance and Sompo, pricing both insurance and financial-market versions of the same weather risk, developing the Australian quanto methodology the desk has traded continuously since 2009, and — on Marty's suggestion — earning a Certificate in Quantitative Finance, mostly, he says, so he'd stop having to take other people's word for how the rest of the market thought about pricing. He also spent roughly a decade sitting through insurance-company audits, external and internal, watching what auditors actually look for. That decade shows up everywhere in ClimateDelta's back office.
The platform itself exists because of a number that offended him. An outside vendor quoted roughly a million and a half dollars to build the system the team needed. A colleague, Ling, looked at that number, decided he could just learn what he needed to learn, and built the first working module himself. Vivek showed it to Marty.
Vivek has been the one steering that build ever since — every module, from underwriting to reinsurance to daily reporting, designed the same way: database tables first, built to last, because recreating a database table is a nightmare and you don't want to go that route. APIs next. The interface last, because the interface is the one part you're always allowed to change your mind about.
One design choice he's particularly proud of: on ClimateDelta, both sides of a transaction see the same valuation. It sounds obvious until you notice how rarely it's true elsewhere — two counterparties, each marking their own book, each somehow showing a profit on the same trade. Vivek built the platform so that can't happen. One transaction, one shared set of marks, both sides looking at the same number.
Ask him what's holding the platform back and he won't tell you it's the technology.
It's the same line, almost verbatim, that comes out of Jonathan's mouth independently, in the same conversation. It's the kind of thing you say when you've spent twenty-five years building something you know works, and the only thing left to prove is that anyone's paying attention.