gustav sparre
// hello (◟ᅇ)◜
i solve problems for a living and it's honestly the best part. there's this rush when a messy data pipeline clicks into place – almost the same feeling watching your team score. data engineering lets me build things that actually matter, where the solution either works or it doesn't, no bullshit.
my work is how i express myself. sounds pretentious but it's true – every pipeline, every model is a reflection of how i think. if it's half-assed, it bothers me the same way a bad sketch would bother an artist.
gothenburg // curious tinkerer // data can't lie ¯\_(ツ)_/¯
things i've built
// the ones i'm most proud of (or at least willing to show)
stuff i've done
// the professional timeline
work
data engineer
2025 – present
volvo cars
went from consultant to full-time. building a modern data platform and making sure stakeholders can actually use their data
read more →
our goal: help the organization identify bottlenecks and make data-driven decisions by building a modern, scalable data platform that stakeholders actually can use.
first mission: kill powerbi.
it had become this franken-dashboard where every stakeholder’s request was bolted on until
nobody could use it anymore.
replaced it with evidence – custom reports per stakeholder, versioned properly, deployed as simple nginx pods.
scales “infinitely"", doesn’t break when someone asks for “just one more filter.”
then the platform. ci/cd was git-flow theater that didn’t actually work. scrapped it for trunk-based development – one golden main branch, everyone deploys and tests without blocking each other. occasionally requires stacked PRs but beats the alternative.
now leading the iceberg migration because our “data lake” needed quotes around it. cleaning happens lakeside, snowflake only touches analytics workloads. costs make sense again.
biggest lesson: ownership means actually owning the problem, not waiting for someone else to fix it.
dagster • dbt • snowflake • spark • iceberg • polaris • kubernetes • evidence • helm
previous gigs →
graduate / data engineer
2023 – 2025
knowit solutions cocreate
started as a graduate, learned what data actually is and the tools to work with it
read more →
kicked off at dun & bradstreet working on pipelines that ingested and cleaned literally all available personal data in sweden. we also maintained and worked on a self-serve BI-ish application where analysts/customers could generate datasets themselves.
then volvo cars. joined a 9-person team doing typical data engineering stuff – pipelines, transformations, the usual. then layoffs hit and suddenly it’s 3 of us responsible for the platform.
baptism by fire, but honestly? best learning experience possible.
side quest: worked on an internal LLM project doing analysis on top of our ETL flows.
also got voluntold to fix the graduate program after finishing it, so i rebuilt it around “learning by doing.” 2024 cohort loved it. 2025 nobody volunteered to improve it again, so i pushed for a global knowledge platform across all knowit offices instead. management actually went for it.
education
msc, complex adaptive systmes (incomplete)
2019 → 2020
chalmers university of technology
left after 6 months to work full-time. turns out you learn more by building stuff - who could have guessed. no regrets.
bsc, mechanical engineering
2015 → 2019
chalmers university of technology
the formulas are long gone, but the concepts stuck. what actually mattered wasn't memorizing how to calculate entropy or sizing bolts – it was learning how to break down problems logically and teach myself whatever i needed to know. turns out that's way more useful than any specific equation.
occasional thoughts
i sometimes write about data engineering, modeling, and how tools actually work under the hood. also whatever's on my mind – ai doomposting, bad leadership, why your architecture is wrong (mine too, probably)
read on substack →