Founding Data Pipeline Engineer

team: product (fashion, taste, personalization)

why this role exists
polopan is building consumer ai where taste, context, and judgment matter more than raw scale.
our models are only as good as the truthfulness of the data beneath them.

this role exists to make sure our catalog and product data pipelines are:

clean
explainable
trustworthy
hard to lie to
we care less about how fast things move, and more about whether they ever need to be questioned again.

what you’ll be responsible for

1. building catalog data pipelines
design and maintain pipelines that ingest, normalize, enrich, and version product/catalog data
define schemas that age well as the product evolves
handle messy, incomplete, and inconsistent data without hiding the mess
make catalog data usable for downstream systems (search, recommendations, personalization)

2. owning data clarity end-to-end
decide what should be logged and what should not
ensure every dataset has a clear purpose and owner
detect and debug silent failures, drift, and data pollution
make pipelines observable, debuggable, and boring in the best way

3. making decisions irreversible
build systems that allow the team to confidently:
trust metrics
kill features
iterate without second-guessing the data
reduce ambiguity for product and machine learning decisions, not add to it

4. setting engineering standards early
establish patterns for data hygiene, versioning, and validation
write documentation that explains why something exists, not just how
push back on over-engineering and under-thinking equally
what we care about (more than speed)

we don’t measure this role by:

number of tickets closed
lines of code written
how fast you ship

we measure it by:

how much confusion disappears after your work exists
how rarely your systems need revisiting
how confidently others can build on top of what you’ve built
sometimes deadlines will exist — not to rush you, but to force clarity on what truly matters.

what we’re looking for

you’ll likely resonate if you:

enjoy turning messy reality into clean, minimal systems
think deeply about schemas, contracts, and downstream consequences
prefer deleting data to hoarding it
care about correctness, not cleverness
are calm under constraint and decisive under deadlines
experience that helps (not all required):

building data pipelines (etl / elt) in production environments
working with catalog, marketplace, or content-heavy datasets
designing event schemas and data contracts
debugging data quality issues that don’t throw errors
familiarity with batch + near-real-time systems
tech stack specifics matter less than your judgment. python would be nice to have.

what this role is ‘not’

not a “ship fast, break things” role
not a model-training or research-heavy ML role
not a growth or analytics-only role
this is a foundational engineering role. What you build early will shape everything that comes after.

how success looks (first 90 days)

we trust our catalog data without caveats
product and ML teams stop asking “is this data right?”
at least one major product decision becomes irreversible because of your work
parts of the system become confidently deletable
if that sounds like a good problem to work on, we’d like to talk.

final note
we’re building this company deliberately.
if you care more about clarity than velocity, and about doing things once, properly, you’ll feel at home here.