Senior Full-Stack Engineer, Research Analytics
Own a whole product: take our research-analytics product for universities to production and grow it from there.
The role
You own our research-analytics product for universities and research institutes, end to end. It turns open scholarly data into an always-current, AI-built view of an institution's research. You take the new version to production and grow it with the institutions that use it.
Why this role
- It is your product: architecture, data, interface and production are all yours.
- Real users early: what you build goes straight to the institutions that use it.
- Rich open data: publications and researcher profiles from sources such as OpenAlex, Crossref, PubMed and ORCID, enriched and scored with LLMs.
What you will do in your first year
- Get the new version through its acceptance tests on real institutional data, with strict isolation between institutions.
- Own the data pipelines and the reports they feed: ingestion, transformation, enrichment and AI-based scoring of research outputs. In the first months a working student runs ingestion and onboarding for new institutions under your lead.
- Onboard institutions and support them as the product grows.
- Build the customer portal and the internal operations console on the product's design system.
- Take the product to production in the cloud, with backups, monitoring and a clean security baseline.
- The stack
- Python, FastAPI and Temporal; MongoDB and MinIO; React with TypeScript, Tailwind, Radix, Storybook and ECharts; Docker and Kubernetes; and whichever AI models are best for each task.
What you bring
- Several years designing, owning and running full-stack products in production. Show us one.
- Python backends and React with TypeScript frontends, both at a professional level.
- Multi-tenant data design, and the discipline it takes.
- Data pipelines and production use of LLM APIs: ingestion from public APIs, structured outputs, cost control.
- Ownership: you have run a system end to end, deployments and incidents included.
- You turn a vague goal into a precise work order an AI coding agent can build from, and you can show us a time you caught an agent's code or tests being wrong.
Nice to have
Temporal; MongoDB; scholarly data sources (OpenAlex, Crossref, ORCID, PubMed); GDPR-compliant use of LLMs; design systems.