Huzzle.com
Domain Expert, Insurance
How your CV stacks up
Upload your CV to see how well it fits this job role
?%
Domain Expert, Insurance
Contract · Fully remote · Huzzle Expert Network · £50-80/hour
About Huzzle
Huzzle builds reinforcement learning environments for frontier AI labs. Our work centres on long-horizon computer use: the multi-step, tool-heavy, judgement-dependent tasks that make up most real professional work and that current models handle poorly.
The role
You will bring working insurance expertise to the design and validation of the environments we ship. You are not annotating data or labelling examples. You are deciding what a realistic task looks like, solving it yourself to establish the reference answer, and judging whether a model's attempt would survive contact with a real file.
Engagements range from a single scoped project to sustained collaboration across a domain, depending on fit and appetite on both sides.
The work suits practitioners who like autonomy, are comfortable with problems that arrive underspecified, and have opinions about what separates competent insurance work from work that merely looks competent.
What you will do
- Define tasks drawn from real insurance workflows, at the length and difficulty a working professional would face, typically spanning many steps rather than a single decision
- Produce the reference solution yourself, including the reasoning, so there is a defensible standard to grade against
- Review model attempts and say precisely where the judgement fails: a coverage reading that does not hold, a reserve that ignores development, a risk accepted outside appetite
- Work with our researchers and engineers to specify the systems, documents, and data a task needs in order to be realistic
- Stress-test tasks for ambiguity, unstated assumptions, and shortcuts that let a model reach the right answer for the wrong reason
- Write clearly and asynchronously, since your reasoning becomes part of what the environment teaches
Reasons to use Rodeo
Start with a chat, not a search bar
Grad scheme, placement, apprenticeship? Not sure what you want yet — that's fine. Your agent talks it through with you and turns "I have no idea" into a shortlist.
It searches the market for you
Every day your agent scans the market matching roles against what actually matters to you, not just keywords on a CV.
Only hits
No noise. No "maybe this fits." Just roles with a clear explanation of why they're right — and where to focus when applying.
Insurance workflows we build
- Underwriting. Risk selection and assessment, appetite and guideline application, pricing and rating, referral judgement.
- Claims. Intake, investigation, and adjudication. Coverage determination. Reserving and loss estimation. Subrogation and recovery.
- Policy administration. Issuance, endorsements, and renewals. Coverage interpretation and wording analysis. Regulatory and compliance handling.
- Portfolio, risk, and actuarial. Loss and combined ratio analysis, exposure management, catastrophe modelling, portfolio-level risk assessment.
We build from how the work is actually done, including the parts that are messy, contested, or resolved by convention rather than rule.
Tools and artifacts
You will work with, or help models reason about, the systems and documents insurance runs on: policy administration and claims platforms, underwriting workbenches and rating engines, loss runs, bordereaux, actuarial exhibits, policy wordings, regulatory filings, and compliance records.
A large part of the job is making tacit expertise explicit. Practitioners know when a submission is wrong before they can articulate why. Getting that from instinct into something a model can be trained and graded on is the hard and interesting part.
Who this suits
- Meaningful hands-on experience in underwriting, claims, policy administration, or actuarial work, typically 5 to 8 years or more
- Credentials such as CPCU, ARM, ACAS or FCAS are welcome but are not a requirement
- Fluency in coverage interpretation, risk selection, and claims resolution, with the ability to explain the reasoning rather than just the conclusion
- A preference for flexible contract work over a permanent role
- Care about precision, and some interest in how AI systems learn to do this kind of work
Get help with your application
Your very own career expert that helps elevate your application to the next level.
If you are strong on the substance and light on the list, apply anyway. We weight demonstrated expertise and quality of reasoning well above credentials.
How engagements work
Fully remote and contract-based. Scope, duration, and weekly commitment vary by project. Contribution and review quality are tracked in the platform, which is what surfaces you for future work across the network.
We are not able to sponsor work authorisation in the United States or elsewhere.
Compensation
Project-based or hourly, set against expertise, scope, and the difficulty of the work.
Typical hourly range: £50-80/hour Highly specialised expertise can go above the range Continuing engagements are usually structured as a monthly retainer rather than hourly
Final terms depend on domain, experience, location, and the shape of the engagement.
Why this work matters
Models are getting good at short tasks and remain unreliable at long ones. Closing that gap needs people who have actually done the work and can say, concretely, what doing it well requires. That is the contribution we are asking for.
“It took my CV and asked me questions relevant to understanding what kind of jobs to suggest for me. Suggestions were almost perfect. Jobs were exactly what I’ve been looking for.”
Jessica, London
Skills
Location