Want to know if this job is worth applying to?
Praha, Česko
Onsite
This role is based in Prague (3 days/week in the office). You build what our clients run on. As a Data Engineer in Keboola Professional Services, you implement data solutions on the Keboola platform for our clients: pipelines, transformations, integrations, data apps and AI use cases that go into production and stay there.
Keboola Professional Services is a small team inside Keboola that builds solutions for new clients and helps existing ones get more out of the platform. You work alongside forward-deployed consultants who own the client relationship and solution design, and who build too. You are the delivery engine: you take a scoped problem, turn it into a working solution, and hand it over in a state the client can own.
You also work in pre-sales. When a deal needs a proof of concept or a deep technical conversation, you build it and present it. Clients meet the people who will later implement their solution, and that is deliberate.
We build agent-first. Coding agents write most of the code. Our engineers own the system around them: the architecture, the environment agents run in (MCP connections, skills, tests and guardrails), several agents working in parallel, and the checks that prove the output is correct before a client ever sees it. Context engineering, meaning capturing a client's data model, metric definitions and business rules in a form agents can use, is as much a part of the job as building pipelines. And more and more of what we deliver is itself for agents: governed data, definitions and skills that let a client's own people hand work to AI.
This role is for someone who gets energy from shipping working things for real clients, and who is comfortable in front of those clients.
Agentic engineering. You build with coding agents every day, not occasionally. You work in tools like Claude Code, connect agents to platforms through MCP, run several tasks in parallel, and know how to break work down so agents can do it. You read the SQL and Python they produce well enough to catch a wrong join or silently dropped rows before the client does, and you know when to take over. Underneath, you know data modelling and have run pipelines in production.
Context engineering. You know that an agent's output is only as good as its context. You write specs, instructions and skills that make agents reliable, and you can capture a messy business (what its numbers mean, who owns them, which rules apply) in a form both people and agents can work from.
Delivery ownership. You've delivered against a scope, a deadline and a budget. You say early when something won't fit, instead of finding out at the end. Ambiguous requirements don't stop you; you ask the right question or make a sensible call and record it.
Client-facing communication. You can explain what you built to a data team and to a finance director. You demo confidently and answer questions directly, and you know when to go deep and when to keep it simple.
English. Professional working proficiency. You'll work with clients across Europe and the US, write documentation and present in English every day.
We tell you up front what good looks like.
Experience: 4+ years in data engineering, analytics engineering or a hands-on technical data role.
Background: You've built data solutions at a consultancy or system integrator, on an in-house data team, or in a customer-facing role at a data vendor. You built the systems; you didn't only consume the reports.
Mindset: You like finishing things. You'd rather ship a working solution and improve it than polish a design nobody runs. You care whether the client is actually using what you built three months later. You treat agents as part of the team: you delegate to them, check their work, and keep improving how they work.