Principal Data Engineer
Founding-level data engineer who owns Polymodal's end-to-end event data pipeline — capture contracts, streaming and batch processing, a BigQuery-class warehouse, and governance/privacy — for its agent-execution platform. Built with Rust, Python, and TypeScript on Google Cloud and Cloudflare.
- Compensation
- Not disclosed
- Location
- Netherlands
- Work setup
- Not specified
- Seniority
- Principal
- Employment
- Not specified
- Experience
- 10+ years
- Category
- Data engineering
- Posted
Skills
Must have
- Uses AI as part of your own data and engineering loop, not as a side tool
- 10+ years building production data systems at a low level, with principal-level ownership of at least one
- Strong in Python/Rust and a polyglot backend; deep in high-volume ingestion and streaming
- Owned quality, reliability, and observability end to end, including SLOs you defined and operated
- Thinks in flows, contracts, and failure modes, not in tables
- Sets technical direction for a team while executing within it
Nice to have
- Have shaped a data function from first hire onward
- Notable GitHub work or hard side projects
Full description
Every role at Polymodal is dual: founding platform builder first, with a data pipelines, governed records, and model-ready signal edge that shapes the platform.
As agents take on more execution, the business needs trusted memory: every captured action flows through one pipeline and becomes the record that humans trust, products query, and our own models learn from. You own that pipeline end to end while it is still early, and this is a charter, not a ticket queue.
Stack: Rust/Python/TypeScript, high-volume event streams, a BigQuery-class warehouse with SQL-defined transformations, Cloudflare and Google Cloud, versioned data contracts, governance. Yours to shape the way forward.
What you'll own
- Push from logs to living signal: make messy and constantly changing event streams queryable, governable, and model-ready.
- Own the capture contracts: versioned schemas that evolve without breaking a downstream consumer.
- Run landing, normalization, and the enrichment lanes that add meaning to raw events: batch and real time, idempotent and replayable, correct as volume grows.
- Keep the warehouse honest: cost, speed, correctness, and the identity-scoped access everything queries through.
- Build the trust layer: quality, lineage, and observability, so anyone can find a record and know exactly where it came from.
- Make the pipeline safe to open up: PII classification, masking, retention, and access control, so new entry points for agents never lower the privacy bar.
About you
- You care about data that is useful, trusted, permissioned, and fast enough to matter.
- You think in flows, contracts, and failure modes, not in tables.
- You set technical direction for a team while executing within it.
Must-haves
- You use AI as part of your own data and engineering loop, not as a side tool.
- 10+ years building production data systems at a low level, not just writing jobs on top, with principal-level ownership of at least one.
- Strong in Python/Rust and a polyglot backend; deep in high-volume ingestion and streaming.
- Owned quality, reliability, and observability end to end, including the SLOs you defined and operated.
Nice-to-haves
- Have shaped a data function from first hire onward.
- Notable GitHub work or hard side projects.
About Polymodal
Polymodal is a funded pre-seed company building the operating layer for agent-executed organizations. We closed our pre-seed in May 2026, backed by Lunar Ventures, Wind Capital, Volve Capital, and early-stage angels. We are early enough for every hire to shape the company, and real enough that the work must survive customers, scale, security, and trust.