Hx Model Developer

"Hx" is shorthand for hurricane, so an Hx model developer builds and maintains the computer models that estimate how much damage and financial loss a hurricane (or related wind and storm-surge event) would cause. The work sits in catastrophe modelling - a corner of the insurance and reinsurance industry where scientists and programmers turn weather and engineering data into loss estimates. The output is used by insurers, reinsurers, brokers and capital markets to decide how much exposure they can take on and what to charge for it.

Starting salary: £30k–£40k
Live graduate jobs: 1 open now
Main way in: Direct entry from a quantitative degree into a catastrophe…
They usually ask for: A strongly quantitative degree

Approximate graduate salary

Typically around £30,000–£40,000 to start, with the upper end more likely for those entering with a PhD or relevant MSc, and London roles generally paying more than elsewhere. This varies a lot by employer type - model vendors, reinsurers and brokers price these roles differently - and bonuses are common in the insurance market.

What you'd actually do

Writing and testing code - commonly Python, R, C++ or SQL - that implements parts of a model: the set of simulated storm tracks, the wind field around the storm, or the 'vulnerability function' that converts wind speed at a location into expected damage to a building
Working with large datasets of historical hurricane tracks, reanalysis weather data, satellite data and insurance claims, and cleaning them into something a model can consume
Calibrating and validating model output against real events - for example checking that the model's simulated losses for a past landfalling hurricane look sensible against what insurers actually paid
Reading hazard and engineering literature (meteorology papers, building-code and structural damage research) and deciding whether a new finding should change the model

Show 3 more

Running regression tests and version comparisons when the model is updated, then documenting exactly what changed and why, because clients who use the model need to explain movements in their numbers to regulators and management
Answering technical questions from model users - catastrophe analysts, underwriters, actuaries - who want to know why a particular region's loss estimate moved, often through written notes or calls
In live hurricane season, contributing to real-time event response work: producing rapid loss estimates while a storm is approaching or just after landfall

How graduates get in

Direct entry from a quantitative degree into a catastrophe modelling vendor (the firms that build and license models) or into the in-house research team of a large reinsurer or broker. This is the most common route and is usually advertised as 'catastrophe model developer', 'hazard researcher' or 'research analyst' rather than using the 'Hx' shorthand.
Via a PhD or MSc in meteorology, atmospheric science, climate science, physical oceanography or structural/wind engineering. For hazard-side roles specifically, postgraduate study is common and sometimes expected, though it is not universal.
Starting as a catastrophe risk analyst at an insurer, reinsurer or broker - running existing models rather than building them - and moving across into development once you understand how models are used commercially. This is a well-trodden sideways route.
Graduate schemes at large reinsurers and broking groups sometimes include a rotation through catastrophe modelling or exposure management, which can lead into development work.

Show 2 more

From a software or data engineering background into the engineering side of model build - the pipelines, performance optimisation and platform work - rather than the science. Employers increasingly recruit for this separately.
Internships and summer placements with model vendors and reinsurers exist but are relatively few; speculative applications and conference contact (the catastrophe risk community in London is small) genuinely matter more here than in larger graduate markets.

What employers ask for

A strongly quantitative degree - meteorology, physics, maths, statistics, engineering, computer science, geophysics, environmental or earth science. Subject does matter here, unlike many graduate roles: you need either the hazard science or the heavy numerical skills.
Typically a 2:1 or above. Some employers are flexible if you have a relevant MSc, PhD or demonstrable coding portfolio.
Programming ability that you can evidence, not just list. Python is the most commonly requested; R, C++, MATLAB, Fortran (still present in legacy hazard code) and SQL all appear depending on the team.
For hazard-focused posts, a postgraduate qualification in atmospheric or climate science is common. For vulnerability (damage) work, structural or wind engineering is often preferred. For platform and pipeline work, neither is required.

Show 2 more

No mandatory professional qualification exists for this role. Some people take actuarial exams (IFoA) or the CII insurance qualifications if they move towards the commercial side, but these are optional and usually employer-funded if taken.
Experience with large or messy datasets - from a dissertation, research project or placement - is often what separates candidates, since much of the job is data wrangling before any modelling happens.

Skills that matter

Scientific programming at production standard

Your code is not a one-off research script - it has to run reliably, be version-controlled and be re-run by someone else years later when a client queries a result.

Statistics, particularly extreme value theory and simulation

Hurricane losses are driven by rare events, so you are constantly reasoning about the tail of a distribution from a short historical record.

Understanding of meteorology or structural damage mechanics

You need to judge whether a model assumption is physically plausible, not just whether the numbers fit.

Data validation and scepticism about sources

Historical hurricane records, exposure databases and claims data are all incomplete and inconsistent, and spotting that early prevents a flawed model release.

Technical writing and explanation to non-scientists

Underwriters and regulators will act on your model's output, and they need the limitations and changes explained without equations.

Commercial awareness of how insurance works

Knowing what a policy deductible, limit or reinsurance treaty does determines how your hazard output is actually turned into a loss figure.

Putting these on a CV? We counted which skills 12,914 live graduate adverts ask for, and what to do with the answer — part of our application advice.

Where it leads

Graduate or junior developer/researcher: working on a defined component of a model under supervision, doing validation, testing and data preparation.
Model developer or senior researcher: owning a hazard or vulnerability component for a region, making calls on methodology, and presenting model changes to internal reviewers and clients. Reaching this point commonly takes a few years but timelines vary widely by employer size and whether you joined with a PhD.
Lead developer or principal scientist: setting the scientific direction for a peril and region, supervising others, and representing the model publicly at industry events. Some people stay on this technical track long term, which is a respected path rather than a dead end.
Alternative moves: into catastrophe risk management at an insurer or reinsurer (using models to set strategy and pricing), into exposure management, into broking analytics, or into climate risk and resilience work outside insurance, where hurricane modelling skills transfer well.
Longer term, some move into head-of-catastrophe-modelling roles, model governance and validation (checking models for regulatory purposes under Solvency II-style requirements), or into climate consultancy and government advisory work.

What people get wrong

“Hurricanes don't affect the UK, so this must be an overseas job.”

London is one of the world's largest hubs for reinsurance and specialty insurance, so a great deal of global hurricane risk is modelled and underwritten here. UK-based teams routinely build models for the Atlantic basin, the Gulf of Mexico, Japan and elsewhere.

“It's a forecasting job - you predict where the next hurricane will go.”

Almost all of the work is about long-run probability, not forecasting. You build a catalogue of tens of thousands of simulated storms that never happened, to estimate how often a given loss level could occur. Real-time forecasting is a small, seasonal part of some roles and absent from others.

“It's a pure science role, so commercial knowledge doesn't matter.”

The model exists to price risk and satisfy regulators. Developers who don't understand how an underwriter or a capital model uses the output tend to build things that are scientifically elegant and commercially unusable.

“You need a PhD to get in.”

It is common on the hazard side and genuinely helps, but plenty of people enter with a good quantitative first degree or MSc, and the engineering, data pipeline and vulnerability sides often value coding and data skills over a research doctorate.

Where this varies

The job looks quite different depending on where you sit. At a catastrophe model vendor - a company whose product is the model itself - you are building and releasing software to external clients, with formal release cycles, documentation and client support. Inside a reinsurer or large insurer, you are more likely to be adjusting or blending vendor models, building in-house views of risk and feeding directly into pricing and capital decisions, with less software-release discipline but closer contact with underwriters. At a broker, the emphasis shifts further towards analytics and client-facing explanation. The balance between hazard science, engineering/vulnerability work and software engineering also varies: some teams split these into separate specialisms, while in smaller teams one person covers all three. Note too that "Hx" is informal industry shorthand - the same job is advertised under many names, so search for catastrophe modelling, cat model developer, hazard researcher and natural catastrophe research as well.

See all 1 live Hx Model Developer jobsFrom employers' own careers sites, updated daily Tailor your CV for this roleFree with an account, 3 a month

General guidance about the role across the UK market, not about any specific employer. Entry routes and requirements vary — always check the individual job advert.