Job Description

About GenAirate

GenAirate builds SnapLine, a submission intelligence platform that converts unstructured broker submissions (emails, PDFs, Excel, images) into structured, decision-ready data for underwriters. SnapLine was built for Lloyd's and the London Market and is now expanding into US specialty insurance. We are SOC 2 Type 2, ISO 27001 and ISO 42001 certified.

The role

You will build and run the product end to end: the interfaces underwriters use, the services behind them, and the cloud infrastructure they run on. A large part of the work is the document processing pipeline that turns messy, inconsistent submissions into reliable structured data. We are looking for someone who cares about how systems are designed, takes responsibility for outcomes rather than tasks, and does not need to be chased.

What you'll do

Design, build and ship full-stack features, from API and data model through to a UI that works for underwriters under time pressure
Build and improve document processing pipelines: ingestion, classification, OCR and parsing, extraction, validation and traceability back to the source document
Design architectures that are clean, efficient and cost-aware, and that hold up as submission volumes and document variety grow
Build event-driven services on AWS and GCP, with sound choices on queues, storage, concurrency and failure handling
Integrate LLM-based components (RAG, retrieval, agentic workflows) with proper evaluation, fallbacks and monitoring
Take ownership of what you ship: deployment, observability, reliability and fixing problems at the root rather than patching symptoms
Raise issues early, make clear technical decisions, and write them down so the wider team can build on them
Work with product and design to turn real underwriter workflows into clear, accessible interfaces
Apply secure development practice as a normal part of the job, in line with our certifications

What we're looking for

3 to 6 years of professional software engineering, with production full-stack experience
Strong Python and TypeScript, with React or a comparable frontend framework
Experience building data or document processing pipelines that handle messy, real-world inputs
Good architectural judgement: you can explain trade-offs in performance, cost, maintainability and reliability, and you simplify rather than over-engineer
Hands-on cloud experience on AWS or GCP, including containers (ECS or Kubernetes), queues or messaging, and infrastructure as code
Solid grounding in APIs, data modelling, and both SQL and NoSQL stores
A track record of owning services in production: debugging, observability, incident response
A self-starter mindset: you take a problem from vague to shipped without close supervision
Clear written communication; you will be working across time zones with a UK and US team

Useful, not required

Experience with LLM applications, vector search or OpenSearch, OCR or document AI tooling
Exposure to Bedrock, Pub/Sub, Firestore or similar services
Insurance or related financial services background. Not needed; we will teach the domain.

What we offer