Senior Product Data Engineer- Data Insights (remote, Europe)

Hey, I'm Tadas . I'm hiring a Senior Product Data Engineer for Data Insights, the team that builds Modash's data products.

Modash helps brands find, understand, and work with creators on Instagram, TikTok, and YouTube. More than 2,700 companies, including Stanley 1913, Sennheiser, and NordVPN, use us to run and grow their creator partnerships.

Data Insights turns raw social media data into the datapoints customers buy: which brands a creator has worked with, how to contact them, where they're based, and which other creators are like them. Most of the work is taking messy public data across 200M+ profiles and making it accurate enough that a brand will act on it.

In this role you'll own datapoints end to end, from the raw signal to what the customer sees, and you'll be judged on their quality.

Why we're hiring

At Modash, the data is the product customers pay for. Brands use our collaboration, contact, and location data to decide which creators to work with, and our largest customers take it in regular bulk deliveries through our API.

We need a senior engineer who can take one of these datapoints from a rough idea to a production pipeline and keep improving it after launch. Right now that means building our own creator-location data, using LLMs to find and validate creator contact emails, and matching sponsored posts to the right brand and its parent company.

You'll join a team of four data engineers who work closely with four backend engineers. You'll make important technical decisions yourself, with strong teammates to back you up, and customers will see what you build.

For a feel of how we build software, read our Engineering Blog .

What you'll own

1. Build the datapoints customers pay for.

You'll turn raw posts, bios, and profile data into signals: brand collaborations, contact details, creator location, related creators, and links between one creator's accounts on different platforms.

2. Make data quality measurable.

Every datapoint we ship has automated quality checks. You'll decide what good looks like for your datapoint, measure it per platform against real coverage numbers, and catch regressions before customers do.

3. Take projects from idea to production.

You'll shape the problem, scope the work, design the architecture, write the code, ship it, and learn from how it performs. We expect senior engineers to own the outcome, even when the problem arrives half-defined.

4. Use LLMs where they pay off.

We run LLM enrichment at large scale. You'll decide where an LLM beats a rule or a simpler model, and work out what it costs across 200M+ profiles before it ships.

What the day-to-day looks like

Most weeks you're on one project for days at a time, with some of your time going to pipeline issues and code review. A typical week might look like this:

• Monday. You spot-check your datapoint's output and find a quality problem. You write a short plan for fixing it and get feedback from another data engineer before you start.
• Tuesday. You build the change in PySpark and test it. It's too slow on the full dataset, so you profile the job and fix it.
• Wednesday. A pipeline failed overnight. You find the cause, ship a fix, and rerun it. In the afternoon you review a teammate's PR, then get back to your project.
• Thursday. You compare results before and after the change on each platform. One number moves more than expected, and you dig in until you can explain it.
• Friday. You merge, watch the first production run, and check the quality reports. You answer a question from a backend engineer about the data, then plan next week.

Every meeting needs a reason, and we protect time for focused work. You'll have a short standup, pair with people when it helps, and get long stretches to plan, build, and ship.

Requirements

What you've done before

• Built data systems at meaningful scale. You have solid experience as a data engineer and have worked with large, complex datasets in production.
• Worked deeply with Spark. PySpark, Scala, or Databricks all count. We use PySpark, but strong Spark fundamentals matter more than the exact flavour. Here is the blog post that helps explain the kind of work we do.
• Turned messy data into signals people rely on. Entity matching, classification, deduplication, or enrichment: you've built reliable output from data that started out incomplete, inconsistent, or wrong.
• Treated data quality as part of the job. You've defined quality checks, tracked coverage and accuracy, and handled a quality drop like a bug.
• Owned full features. You've taken a feature through planning, scoping, architecture, implementation, release, and iteration.
• Written production Python and SQL. You care about code quality, system design, and maintainability.
• Worked with orchestration and cloud infrastructure. Experience with Airflow or AWS Step Functions, and with services such as EMR, Glue, Athena, DynamoDB, S3, Kinesis, Lambda, or ECS, will help you get moving quickly.
• …