Machine Learning Engineer — One Matching
Team: One Matching · Location: Prague / Czech Republic (hybrid) · Level: Mid-level, with room to grow
About the team
One Matching builds the systems that connect millions of shop offers to the right products across our marketplaces (Czech Republic and Slovakia). Our Offer-to-Product matching system uses search, machine learning, and simple, practical engineering to make large product data useful. We care about results that our content and product teams can feel — how many offers we match, how correct the matches are, how much we cover.
What we mean by "ML Engineer"
For us, an ML engineer works across several areas, not just one: software engineering, training models, verifying models (including how sure we are about the results), and data analysis. You do not need to be the best at every one, but you should be comfortable with all of them and able to move between them as the task needs.
What you will do
- Own problems from start to end — understand the business need, choose an approach, ship it, and measure it in production.
- Decide if a problem even needs machine learning. Often a good rule or a simple baseline is the right answer, and we want you to start with the simplest thing that gives value.
- Build and improve matching, ranking, and data-quality solutions — using classic ML, search, and more and more LLM-based methods where they help.
- Work closely with content and product teams: turn unclear business problems into clear, planned steps, and explain the trade-offs.
- Move decisions forward — suggest a direction, write it down, and help the team reach a conclusion.
- Look at the whole system: understand how the parts (search, features, model, thresholds, deployment) and the nearby services work together.
What we are looking for
- Practical, business-first thinking: you see technology as a way to reach a goal, and you weigh effort against value.
- Real experience with applied machine learning: training and checking classic ML models (for example XGBoost, scikit-learn), thinking about uncertainty, doing error analysis, and knowing when ML is needed.
- MLOps: comfortable putting models in production and operating them (deploy, monitor, maintain).
- Good software engineering: you write reliable production Python and can own your code from start to end. Cloud experience, ideally GCP.
- Systems thinking and planning: you can define the work, put the steps in order, and give realistic estimates.
- Ownership and action: you can make a decision with incomplete information and improve it step by step.
- Clear communication with both engineers and non-technical people.
Nice to have
- Search experience (Elasticsearch or similar).
- Vertex AI (our ML pipeline platform).
- Experience using LLMs in production (annotation, evaluation, extraction).
- Czech or Slovak language, or experience with the local e-commerce market.
Our stack
Python, Google Cloud (GCP), Elasticsearch, XGBoost are the base of what we build, together with classic ML and LLMs. We use LLMs in production today, and we are starting to do more agent-based development. We expect this area to grow in the next months.
How we work
Small team with a lot of ownership. We prefer to ship and learn instead of long analysis. We keep solutions as simple as the problem allows. We expect everyone to help move decisions forward. We use modern AI development tools every day.
What we offer
- Real Impact: Your code will shape the shopping experience for millions of people.
- Hackathons & Conferences: We’ll support your participation in events like DevFest or devopsdays.
- Flexibility: Home office (up to 3x a week) and flexible working hours are a standard for us, not just a perk. That said, we still believe in the power of face-to-face connection—for team collaboration and scrum ceremonies, we all meet in the office at least once every two weeks.
- Career Journey: We have a clear roadmap to help you grow and move forward.
- The Classics: 25 days of vacation, extra time off, a Cafeteria benefit system, and a Multisport card.
