Work on real AI, on real systems, with real customers.
Datablooz is always open to senior engineers, ML specialists, and delivery leads who want to ship production AI rather than decks about production AI. Below is the shape of how we work and what we are hiring for.
Four things that make working here different.
We are a small, senior team. We do not hire juniors to absorb headcount. Everyone ships something.
Real production work, every week.
No disposable POCs. You ship systems that get operated, monitored, and retrained in real customer environments.
Named accountability, not anonymous throughput.
Your name is on the engagement. The customer knows who delivers their program. Your work is visible.
Governance as part of the craft.
Eval harnesses, drift monitoring, human-in-the-loop. You learn to build AI that survives an internal audit.
Senior peers, small team, clear ownership.
Flat structure. Decisions get made in the room. You work alongside people who have shipped dozens of production AI systems.
We are hiring across engineering and delivery.
- Senior ML engineer
End-to-end ownership of modelling, evaluation, and production deployment. Comfortable with eval harnesses, drift monitoring, and post-launch operation.
Belgrade or remote (EU hours) - Senior data engineer
Data platforms, pipelines, feature stores. You treat data as product, not plumbing. You leave environments better than you found them.
Belgrade or remote (EU hours) - Program lead
You run Blueprint to Scale engagements. The customer trusts you. The team follows your operating rhythm. You say no when the scope does not fit.
Belgrade or hybrid - Open application
Not seeing your role? If you have shipped production AI and want to work this way, write to us. We hire when we meet the right person.
Anywhere in the EU and SEE region
Applying is an email, not a form. Reach out with your CV, GitHub or portfolio, and a short note about one system you shipped that you are proud of.
Email info@datablooz.com.
We read every application. Expect a reply within one week. The process is typically three conversations and a paid technical exercise.