> Markdown version of [/jobs/ext/3240598-staff-backend-engineer-platform](https://www.wearedevelopers.com/jobs/ext/3240598-staff-backend-engineer-platform). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Staff Backend Engineer, Platform

- **Company:** Zeit AI (YC S24)
- **Location:** London, UK
- **Salary:** £95,000.0 - £270,000.0
- **Contract:** Permanent contract
- **Skills:** Databases, Data Infrastructure, Shard (Database Architecture), Virtual Private Networks (VPN), PostgreSQL, TypeScript, WebSocket, Snowflake, Caching, Backend, Production Code, Cloudflare, Data Management, Data Pipelines, Docker, Databricks
- **Published:** September 21, 2026
- **Apply:** https://www.indeed.com/viewjob?jk=5ecf469f167e8e3e

## About the Role

* staff or principal level, still hands-on, writing production code
 * built and scaled multi-tenant data platforms in a B2B startup or platform company before
 * deep Postgres or database knowledge: query plans, locking, I/O, caching, replication, sharding
 * understands startup pace and ownership without specs
 * pragmatic over ideal, wants to build foundations right without overengineering
 * strong TypeScript
 * English C1, German a plus

Nice to have

 * experience at a data platform or warehouse company (Snowflake, Databricks, Cloudflare or similar)
 * networking depth: proxies, VPN, routing

## Description

* ZeitMind is an autonomous data engineer for mid-market businesses. Agents and people work the same company data: ingest, pipeline, clean, analyse, build operational apps
 * we are growing fast and will be onboarding multiple enterprise customers a week. The backend isn't ready yet, your impact will decide the speed at which we can grow
 * load growth is compounding: more customers, multiple users per customer, many agent interactions per user and multiple agent artifacts (data pipelines, schedules, analyses) coming out of the interactions
 * you own the platform

What you'll do

 * scale a multi-tenant Postgres under unpredictable load: query and I/O optimisation, read/write separation, horizontal scaling
 * make agent-driven workloads survivable: one user can spin up 30 table syncs, some running hours, while others query large sources
 * manage the sandbox layer: Docker containers where agents run queries and reverse-engineer files, with compute and memory budgets
 * own connectivity into customer networks: SOCKS5 over WebSockets, authentication, routing
 * scope the architecture and build it