Data engineer - Vessel data platform

Cetasol; Göteborgs kommun; On-site; SEK 900,000 - 1,300,000; Full time

Job summary

Cetasol is seeking a data engineer to own the data path from the onboard main unit to the cloud dashboard. You will build validation, diagnose data defects, and model the warehouse layer used by reporting and ML.

The role requires 5+ years of production data system experience with Python, strong time-series data skills, and experience with intermittent connectivity. Location is Gothenburg, on-site, with English as working language.

Qualifications

• 5+ years building production data systems with Python as primary language.
• Real experience with time-series or streaming data, irregular sampling, gaps, backfills.
• Strong SQL and warehouse or dimensional modelling.
• You test data, not only code: assertions, contracts and CI-based testing.
• Comfort with systems that must function with intermittent connectivity.
• Ability to sit with a domain expert and translate needs into specs.

Responsibilities

• Own the data path from onboard unit to cloud dashboard: ingestion, normalisation, aggregation into trips/legs/voyages.
• Build validation into the data path to catch wrong values before customers see them.
• Make data defects diagnosable with lineage and tooling for faster tracing.
• Model the warehouse layer that supports reporting, analytics and ML.
• Shape how we prevent recurrence by owning the regression story for the data layer with QA.

Skills

Python

Time-series data

SQL

Data warehousing

CI data testing

Intermittent connectivity

Domain expert collaboration

Tools

AWS

ETL tools

Job description

Cetasol builds iHelm, an AI-driven energy-optimisation and decision-support platform for commercial vessel fleets. An onboard main unit runs digital-twin models at the edge — including through weeks at sea without connectivity — a Captain's Display gives the bridge live fuel and speed guidance, and the iHelm cloud dashboard gives fleet managers performance analytics plus automated EU MRV and IMO DCS reporting. Operators using it report fuel savings of 10–25%. We are based in Gothenburg, backed by BackingMinds, Shift4Good and Sarsia, and deliberately focused on small and mid-sized vessels — the segment the large vendors overlook.

About the role

• Own the data path from the onboard main unit to the cloud dashboard — ingestion, normalisation, and aggregation into trips, legs and voyages.
• Build validation into that path so wrong values are caught before a customer sees them: unit and conversion checks, sensor plausibility bounds, drift and gap detection, and reconciliation between views that report the same metric.
• Make data defects diagnosable. Today, "this number looks wrong" can take weeks to trace. We want lineage and tooling that make it an afternoon.
• Model the warehouse layer that our reporting, analytics and ML work is built on.
• Shape how we prevent recurrence instead of re-fixing: own the regression story for the data layer together with QA.

What we are looking for

• 5+ years building production data systems, with Python as your primary language.
• Real experience with time-series or streaming data — irregular sampling, gaps, late-arriving records, backfills.
• Strong SQL and warehouse or dimensional modelling.
• You test data, not only code: assertions, contracts and expectations that run in CI.
• Comfort with physical measurement. You know why an instantaneous fuel-rate signal and a voyage consumption total can disagree, and how to establish which one is lying.
• Experience with systems that must behave correctly over intermittent connectivity.
• You can sit with a domain expert, understand what they mean, and turn it into a specification.

Nice to have

• Maritime, automotive, energy or industrial IoT background.
• AWS.
• Exposure to digital-twin or physics-based modelling work.
• Experience being the first senior person in a data function.

Practicalities

• Location: Gothenburg, on-site.
• Working language is English; Swedish is not required.