Co-Founder, Reinvent Catastrophe Modelling

WHAT WE'RE WORKING ON AT DSV

Catastrophe modelling relates natural disasters to future payouts. With the advent of climate change, however, nature is no longer behaving like it used to, and past data is becoming less predictive of future events. We also have less historic data than we’ve ever had – emerging environmental phenomenon such as wildfires are thus named because they haven’t happened at this scale of loss before. Many are trying to solve this problem by harvesting ever more data on the physical characteristics of buildings through satellite imagery and LIDAR technology. We believe this methodology is fundamentally flawed: the best indicators of housing integrity are not visible externally (and thus not obtainable through satellite imagery, no matter the resolution quality) whilst deteriorating building quality is not what is driving increasing insured losses – changes in urbanisation and socioeconomic factors are. We are looking to revolutionalise catastrophe modelling by using machine learning, big data and cutting-edge demographic research to build new methods for loss estimation based on a deep understanding of the urban environment rather than purely physical building characteristics.

In order to achieve this, we’re looking to develop:

• efficient and robust ETL processes

• advanced feature extraction capabilities (machine learning, computer vision, OCR)

• APIs for connection to data sources such as Google Street View, NASA satellite imagery, infrastructure documents, city planning documents etc.

• back-end compatibility with exiting open source catastrophe modelling frameworks such as Oasis

• a front-end for interaction with our loss estimation modules

• innovative vulnerability curve methodologies

THE OPPORTUNITY FOR YOU

Joining DSV is a fantastic opportunity to develop concepts for and launch your own startup. Working with us, our sector-specific advisors and our current founders, you’ll have access to a completely new kind of environment for creating science-driven companies, one that does not start with university IP, but with a systematic rework of an opportunity area from first principles.

Given the current stage of the project, we expect less than 3 months until incorporation and spin-out to start the early tech development / proof-of-concept work with investment secured from DSV, and other funding sources.

WHO SHOULD APPLY

You are motivated to build a company that benefits the planet. You are collaborative, committed to your teammates, and able to communicate clearly. You are creative, always asking why, and actively enjoy building a network around the problems and opportunities that you go after. You bring broad technical knowledge that is relevant to the opportunity area (see below).

Apply if you have expertise in:

• end-to-end software development experience (C/C++; Java; Python; SQL; JavaScript; React)

• significant back-end data engineering experience, having ideally developed APIs and ETLs in the past

• experience working with big data (Apache Spark, PySpark, Kubernetes, Docker, AWS, Azure, DynamoDB)

• experience building ML/AI/Computer vision models and undertaking advanced feature extraction from GIS and/or text documents

ideally, combined with knowledge in any of the following:

• experience with PostGIS

• familiarity with open data formats i.e. OSM

• knowledge of ETL scheduling

• knowledge about catastrophe modelling and vulnerability curves

• experience in insurance / insurance risk modelling