What it's all about

The Team:

We are building the analytical backbone of a company that believes decisions should be powered by clarity, not guesswork. Our Business Intelligence team builds on top of a data platform that has to be there every morning - healthy, current, and trusted.

Making that happen takes a core data engineer who owns the operational reliability of our platform end-to-end. Someone the rest of the team can count on to keep the lights on, catch issues before they become incidents, and push the platform forward with us rather than just holding it in place.

Your Role:

This is a core data engineering role. You own the operational health of our data platform, and you keep making it better.

Your working hours - 5am-2pm CEST - put you in a timezone that naturally overlaps with our overnight processing window. That means overnight maintenance, failed jobs, and quality incidents land inside your normal working day, not at 5am in your bed. You catch them, fix them, and hand over a healthy platform before the European team logs on.

We're looking for someone experienced enough to operate independently. You don't need a ticket telling you something is broken - you can read logs, trace through SQL and Python, and figure out what happened. You care about data quality as a craft, and you have the confidence to walk up to a data owner and say "this feed is wrong, here's why, and here's what we should do about it."

This isn't a greenfield architecture role. The platform exists, but it is a long way from finished. Roughly half the job is keeping it healthy; the other half is leaving it better than you found it - new source domains modelled properly, quality gates where there are none today, slow queries made fast. Keeping the lights on is the floor here, not the ceiling.

How We Work:

A large share of what we build is AI-assisted, and some of it is AI-generated. Claude Code, MCP servers, and LLM tooling are part of the daily toolchain across BI, internal tooling, and data pre-processing. Everything lives in Git, ships through GitLab CI/CD, and gets reviewed.

That only works because someone puts the rigour in behind it. Generated SQL still has to survive an execution plan. A pipeline an agent wrote still has to be correct at 4am when a source system quietly changes shape. This role is that layer: you will use these tools heavily, and you will be the person who checks what comes out of them - reading the query instead of trusting it, validating numbers against the source, catching the plausible-looking answer that is wrong.

So we need someone fluent with AI tooling and unwilling to take its output on faith. Those two things aren't in tension here. Together they are the job.

What You'll Be Doing

Keep it running:

• Monitor overnight processing, resolve failures, and hand a healthy platform to the European team each morning - you are the BI team's first line of operational defense during their off-hours
• Keep the orchestration layer healthy: Airflow DAGs and the Python jobs behind them across Windows and Linux VMs - failed tasks, backfills, and dependencies that match how the data actually flows
• Investigate and fix recurring issues in our on-prem Microsoft SQL Server environment, including cross-system access through linked servers, OPENQUERY, and PolyBase
• Hunt down data quality gaps - stale feeds, broken joins, silently-changing source systems - and drive them to resolution with the data owner

Keep making it better:

• Extend the platform as the business grows: model new source domains into the warehouse and build data models analysts can use without needing a translator
• Build automated data quality gates - freshness, volume, referential integrity, business rules - so bad data fails loudly at the door instead of surfacing in a dashboard three days later
• Tune slow SQL - queries, stored procedures, indexes, execution plans - so the platform gets faster, not slower, as the company grows
• Turn recurring fixes into permanent ones through better alerting, logging, runbooks, and automation, so the same incident stops coming back
• Be the last check on AI-assisted work before it reaches production - review generated SQL and pipeline code, and build the tests that let the rest of the team move fast on top of it

What You'll Need

• 3+ years of core data engineering experience in a production environment
• Microsoft SQL Server professional - stellar T-SQL plus real optimization depth (indexing strategies, execution plans, query tuning, partitioning), the instinct for which of those a slow query actually needs, and comfort reaching across system boundaries with linked servers, OPENQUERY, and PolyBase
• Data modelling judgment - you can design warehouse tables and dimensional models that hold up as sources change and analysts ask new questions, and you know where to put a quality gate so it catches problems instead of generating noise
• Strong Python and production Airflow - in-depth Python for pipelines, transformation, and automation, and DAGs you have authored, operated, and debugged for real: scheduling, retries, backfills, dependencies, and tracing why a task failed rather than just clearing it
• AI-assisted engineering, and the rigour to verify it - hands-on with Claude Code or a comparable agentic coding tool, MCP servers, and LLM-powered workflows, paired with the habit of checking what they produce: reading the generated SQL, looking at the plan, validating output against the source. Both halves are must-haves here, not differentiators
• Git and GitLab - branching, merge requests, code review, and CI/CD pipelines as everyday habits
• Environment fluency - you can debug on both Windows and Linux VMs, and you have working knowledge of at least one major cloud, Azure or AWS
• Ownership by default - you chase why something broke instead of restarting it, you carry issues end-to-end without being pointed at them, and your English is clear enough to tell a data owner their feed is wrong and be taken seriously
• Availability to work consistently within the 5am-2pm CEST window, in a timezone where these are normal daytime hours (roughly GMT+4 to GMT+8)

Nice If You Have

• Microsoft Fabric experience - Lakehouse or Warehouse, Data Factory pipelines, OneLake. Our platform is on-prem today and Fabric is a direction we are actively exploring
• Experience building your own MCP servers or internal LLM tooling, rather than only consuming them
• Familiarity with Power BI or another BI tool as a consumption layer
• Exposure to fast-growing B2C or subscription-driven businesses

What You Can Expect On Board

• Remote-first collaboration across a truly global team, with EU meet-ups and an annual company summer getaway
• Clear, predictable working hours - this role is designed around 5am-2pm CEST, not on top of it
• Real ownership of the operational layer - you run it, and we trust you to run it
• A team that builds with AI in real production use - Claude Code, MCP servers, LLM tooling - and expects you to say so when the output isn't good enough
• A culture that values clarity, structure, and long-term thinking over quick fixes
• Close collaboration with BI, analytics, and data-owning teams across the company

If you're the kind of engineer who takes pride in a system running smoothly, who chases a flaky pipeline until it's actually fixed, and who wants to hand back a platform measurably better than the one you inherited - this is your seat.

You keep the platform healthy, and you keep making it better. That's the deal.