[Remoto] Desenvolvedor(a) Full Stack Sênior — AI & Agents na WeatherWise | 4 dias / 32h

### About the company WeatherWise (Ehabitation Ltd, trading as EHAB) is a London-based weather risk specialist helping construction, energy, agriculture and logistics businesses understand, plan for and insure against weather risk. We run WeatherWise Construction (weather-delay analysis, Monte Carlo simulation and schedule risk) and WeatherWise Studio (an AI-native parametric insurance platform). Our agent, Kite, helps brokers discover relevant weather events, structure and price cover, generate client materials and route submissions to suitable carriers. A weather-data engine drawing on 180,000+ stations and global datasets powers both platforms. Our team is small and senior. Trust, auditability and correctness matter. This position works permanently from Brazil. ### Company website https://www.ehab.co/ ### Job description **Semana de trabalho de 4 dias — 32 horas semanais. 100% remoto do Brasil. Contratação PJ. Remuneração mensal: R$ 15.000 a R$ 25.000, conforme experiência.** **Senior Full-stack Engineer (AI & Agents), reporting to the CTO.** Own features end to end, with a strong backend focus: data models, services, agent tools and the React interfaces they drive. - Design, build and run APIs, data pipelines, pricing and simulation engines, and integrations for WeatherWise Studio and Construction. - Extend agent tools, orchestration, memory, evaluations and observability. Build reliable workflows with guardrails, graceful failure and human approval steps. - Own weather-data ingestion, storage, indexing and fast querying of decades of time-series and geospatial history. - Ship React/TypeScript interfaces for your features, working with our design-led front-end engineer. - Build clean, documented, versioned APIs for brokers and carriers. - Own CI/CD, infrastructure as code, monitoring, security and incident response. - Use AI tooling daily and help the team improve how it builds. Work directly with the CTO, with substantial ownership of architecture and technical direction. Agents are both part of our product and how we build it. We prototype quickly, talk to customers and own outcomes. ### Requirements **Required** - 6+ years building and running production systems, with ownership of backend, infrastructure and front end. - Strong Python and TypeScript, including the ability to ship React interfaces. - Hands-on production experience with LLM-powered or agentic systems: tool use, orchestration, prompt/context design, evaluation and failure handling. - Solid data engineering: pipelines, time-series and geospatial data at scale. - AWS or GCP, containers, infrastructure as code, CI/CD and observability. - Daily use of AI coding tools such as Claude Code or Cursor. - Ability to scope ambiguous problems, make pragmatic decisions and ship in a small senior team. **Nice to have** Weather/climate data, NetCDF/Zarr, ERA5, pricing or simulation models, insurance/fintech, public APIs and LLM evaluation frameworks. We care far more about what you have shipped than which exact tools you used. ### Benefits Four-day working week: 32 hours per week. Fully remote from Brazil. PJ contract. Direct access to the CTO and substantial ownership of technical decisions. ### Salary range BRL 15,000–25,000 per month (R$ 15.000 a R$ 25.000/mês), depending on experience ### Level Senior ### Contract type PJ ### Work model Remote ### Location Brasil — 100% remoto ### How to apply Email [email protected] with the subject "Senior Full-stack Engineer". Skip the cover letter. Send two or three things you have built that you are proud of (links, a repo or a write-up), plus a couple of sentences on what you would change about them now. If you have built something with agents, lead with that. We will come back to every application and aim to run the process from first conversation to offer in under three weeks. ### Confirmations - [x] This issue contains a single job opening. - [x] The job is backend (or fullstack with a backend focus). - [x] This is not an AI training, data annotation or model evaluation gig.