Data Engineer

Lionsville

Den Haag

On-site

EUR 51,000 - 65,000

Full time

3 days ago

Be an early applicant

Application generator

Don’t send a generic resume — generate a resume and cover letter tailored to this exact role.

Get past ATS filters

Benefits offered by this job

Electric lease car

MacBook or Windows laptop

Claude Code premium seat

25 vacation days and 8% holiday pay

€4,000 training budget per year

Job summary

Lionsville is hiring a Data Engineer to own end-to-end data pipelines from operational systems into analytics platforms for maritime, logistics, and critical infrastructure. You will define data contracts and ensure data quality while enabling AI-driven improvements.

Work in a customer-facing environment, focusing on real data, timeliness, and reliable feeds. The role requires strong data engineering skills, SQL fluency, and comfort with hybrid tech stacks in a European context.

Qualifications

Ownership, a bias to solutions, and clear communication across organisational boundaries.
Strong architecture capability for data platforms — ingestion, storage and scheduling.
Deep data engineering skills in ETL/ELT, orchestration and scheduling; fluent SQL and streaming experience.
Experience working with AI in a practical, production-ready way.
Complex data-model design with data contracts and schema versioning.
A proven sense for data quality, handling staleness and drift.
Willingness to travel to customer sites; Randstad resident; English required; Dutch welcome.

Responsibilities

Own the data path from operational systems to analytics platforms across multiple customers.
Define and enforce data contracts, data freshness and source trust.
Build and operate pipelines with batch and streaming, orchestration and backfills.
Ensure data quality, monitoring and observability across feeds.

Skills

Data engineering

ETL/ELT

SQL

Streaming

Data contracts

Communication

Tools

NoSQL

Batch/Streaming orchestration

Job description

Let your career roar as a Data Engineer at Lionsville

Somebody has to make the data trustworthy before anything intelligent can be built on top of it. At Lionsville that is you — on systems where a wrong number is not a bad chart but a bad decision, taken live. Short lines to the people who decide, so what you find in the data changes the plan rather than filling a report.

What will you do?

You own the path from operational systems into platforms people act on — in maritime, logistics, industry and critical infrastructure, and across more than one customer and domain over time. The data is real, imperfect and arrives at awkward times. Making it dependable is the job.

You are an AI maestro

Writing the code is largely your agent's job. Yours is making sure the specs carry what the customer actually asked for, and that the agent has the context and the skills to deliver it. If you run out of tokens, ask for more — we would rather pay for the best tools than watch you type.

You turn feeds into contracts

You write the first data contract against real payloads rather than assumptions — event streams, scheduled extracts, database views and, yes, spreadsheets. You define what each field means, how fresh it is, where it came from, and what the platform is allowed to conclude from it. You measure before you model. Before anyone tunes an algorithm, you establish what the data actually says: volume by segment and time window, throughput, completeness, the gaps nobody mentioned. Your assessments become the baseline the whole plan is set against, and you are the one who says so in the room.

You build the pipelines

Extraction from systems that were never designed to be read from, transformation into models people can reason about, and orchestration that runs unattended. Incremental loads, idempotent reprocessing, backfills, and late or out-of-order data handled deliberately rather than discovered in production.
You work across diverse stacks doing it: batch and streaming, SQL and NoSQL, warehouses and lakes, cloud-agnostic architectures — the major providers, and equally on-premises and European ones — with systems integrated through APIs and event-driven solutions. Which of those a project uses is a decision rather than a default, and you will be part of making it.

You make quality a first-class feature

A stale feed must be visible as stale, never as current. A missing critical input must stop a recommendation rather than degrade it quietly. You build the checks that make those rules real, the monitoring that shows the state of every feed at a glance, and you hold that bar when a deadline argues against it.

About Lionsville

We build software that runs real operations — maritime navigation, logistics hubs, critical infrastructure. A business tech boutique: we take on enterprise-critical problems from strategy through to execution, in small senior teams close to the customer. We work as a pride: we back each other, we say what we think, and we finish what we start.

What you bring to the pride

Ownership, a bias to solutions, and clear communication. You work across organisational boundaries — the source systems are usually somebody else's — so you ask precise questions and you are easy to say yes to

Strong architecture capability for data platforms — ingestion, storage and scheduling designed to be understood and changed by someone else, with the trade-offs written down rather than implied

Deep data engineering — ETL and ELT, orchestration and scheduling, with SQL you are genuinely fluent in and real experience of streaming as well as batch. With it the judgement that only comes from having been burned: idempotency and replay, schema drift, and what a backfill actually costs. The tooling matters less than the depth

Experience working agentically with AI — not curiosity about it, but a way of working you already have, and opinions about where it breaks down

Complex data-model design, across SQL and NoSQL, with data contracts and schema versioning — and the discipline to treat a contract as a promise

A real feel for data quality — staleness, completeness, silent drift. The curiosity to ask why a fifth of the records are missing a field everyone assumed was always populated, instead of filtering them out, and the honesty to say \"we cannot conclude that from this data yet\"