Company / Careers / Engineering
Economic Data Scientist
The dataset is the economy itself: 693M+ companies live, and largely unexplored. Most careers in data run on a sample, a slice of someone else's collection; here the whole population is the daily working material. The role on top of it has no clean name. Part forward deployed engineer: proof-of-concepts built for a live prospect, on their timeline, iterated until they convince. Part quant: hunting alpha in the raw, mostly in our own data, occasionally in a new source you'll integrate into the graph. Part applied data scientist, where applied is literal: hands in the data every day, cleaning it, checking it, rebuilding on top of it, until it serves the purpose. Full ownership throughout, you take the questions clients bring, dig for the answer, and defend what you find.
The way we work is not for everyone. Problems don't come broken down into bite-sized pieces: you'll get real ownership early, sometimes before you feel ready, and from there it's yours to run with. The pace is sustained, and with the company growing exponentially, it won't ease up any time soon.
Better to know in five minutes than to find out in month three. If reading this feels like a no, trust that feeling. If it raises flags, bring them up early, and we'll give you straight answers. If autonomy and high stakes are what you're after, you may have just found the right place to grow.
Who we are
- A deeptech startup of 60 people in Bucharest, building the machine that reads the company world.
- Our customers include the giants of the data world, the same companies you'd expect to be our competitors.
- Founded in 2019, VC-backed. The team stays small on purpose, so decisions stay with the person doing the work.
- Most of the team is technical, founders included. The decisions that shape the company are made by people who understand the low-level work behind it.
- What we run in-house (from our crawlers to our own LLMs) is what several entire companies get built around. Here it's only one team.
Who you are
- You own your work like a founder owns a product.
- A growth mindset, able to capitalize on unprecedented contexts through your skills and abilities.
- A strong problem solver, visible in the way you deal with the tension between brief and shipping.
- Resilient, especially in front of failure, the kind that always comes paired with pioneering work.
- An appetite to grapple with a variety of technical challenges.
What we do
- Crawl the entire internet and decide, page by page, what matters.
- Interpret everything we find with LLMs we train ourselves, in 125 languages.
- Resolve every source under the company it belongs to, in one living graph of 186M+ companies.
- Trace every datapoint back to the source it came from, so every judgment can be defended.
- Keep the history, so we notice when any one of them changes.
What you'll do
- Build proof-of-concepts around a live prospect's case: assemble our data and tools into something they can hold, and iterate, ten or thirty times, until it wins.
- Hunt through a firm-level dataset of 693M+ companies for the pattern that answers a prospect's actual question: the insight nobody has looked for, that lands the deal.
- Turn a client's question into an analysis: frame the problem, choose the method, iterate until it convinces, and defend the conclusion out loud, in front of the client.
- Treat cleaning and validation as first-class work: the data is real, straight from the web, and yours to make dependable.
- Ship on client timelines: requests come in fast, answers go out in days, and you see them put to use.
- Work the full range of the seat: quality deep-dives, new sources worth bringing into the graph, findings that become published stories.
What you'll find here
- Founders who built products serving billions of users, across web, big data, and cloud.
- Infrastructure most companies rent, built and run in-house: crawlers, GPUs, custom LLMs.
- Scale you'd otherwise wait a career to touch: 4.2B pages a month, 3M tokens a second, from day one.
- Problems no one has solved before, some so far out that academic papers are the only formal guide we have.
- Five disciplines under one roof (crawling, ML, big data, infrastructure, product) close enough that you see how all of them actually work.
What we look for
- Statistical grounding solid enough to tell a real correlation from noise, and to defend the difference.
- The instinct to interrogate the data before trusting it, and the pace to ship a solid analysis this week rather than a perfect one next month.
- Comfort processing genuinely large datasets, tools-agnostic: SQL, Python, R, Scala, Spark, Excel, whatever gets it done.
- The resilience to iterate dozens of times on the same problem, including the restarts a client's request forces, without losing the thread or the pace.
- Comfort being the technical person in the room with a client: answering live, and saying "I'll check" when that's the true answer.
- Real curiosity about how businesses and markets behave, whether you got here through economics, computer science, or somewhere else entirely.
What we offer
- Fair, market-aligned base, ESOP tied to impact (for permanent roles), and a high salary growth rate tied to performance.
- Decisions that would need three approvals elsewhere are yours to make, and yours to answer for.
- Growth at the pace you can take, and responsibility that expands as fast as you prove you can hold it.
- What you build lands in front of some of the largest companies in the world, often within weeks.
- Steep learning curve, no matter how experienced you are, with people who've climbed it a desk away when you're stuck.
What we expect in return
- High tolerance for ambiguity, marked by your ambition to push forward with incomplete information.
- High speed and uncompromising quality in your work.
- The ability to quickly and effectively evaluate technical tradeoffs and translate them into relevant scenarios
- Genuine aversion to any customer or colleague struggling with something you delivered.
- Ask what would make it ten times better, starting with your own work.
How we hire
A process built to help both of us decide.
We keep it informal because that's how we actually work. And we take it seriously because one hire can change a company this size. The way we communicate throughout is designed so that either of us can openly say it isn't working, even mid-interview.
- 1Home AssignmentA hands-on problem cut from the real work of this role. You'll know what the job is like before you say yes to it, and so will we.
- 2Team interviewsSessions with the people you would actually work beside, one of which is face-to-face and includes live tasks.
- 3DecisionA yes or a no, typically fast. We do not ghost.
Careers
Recognize yourself in all of this?
Send your application and a note on why this role fits. We read every one, and if it matches what we're looking for, you'll hear from us.
Adjacent seats