About the Role
We are looking for a data scientist who wants to not just build models, but to see their real effect on the business. You will develop models for one of the key areas, and the primary client will be the portfolio manager – meaning decisions will directly impact portfolio management and business results.
We don't create "shelfware": every model has a measurable business effect, and you will see how your work is reflected in the company's numbers.
Specificity of the Area
The business is based on several independent brands, each with its own product line. Some client connections are already described in the Neo4j graph database, but there is huge room for improvement here – there is space to grow and enhance.
In addition to building models, you will be involved in research ad-hocs: for example, evaluating the potential of new variables or data sources before they go into production.
How the Modeling Pipeline Works
- Task Definition - together with the client, you define the model's goal: what does a "good model" mean in this context and what business problems should it solve.
- Data Handling
- for some sources, variables are already prepared;
- complex new sources are implemented by the data engineering team (DE), and you act as the task setter as an internal client;
- you will work on some sources at a basic level independently.
- Modeling - experiments, selection, and justification of the best model.
- Deployment - implementing the model as a class according to a template in MLflow.
- Monitoring - setting up the model for monitoring through the company's internal service.
What You Will Get
- A full cycle from idea to production and monitoring – with no gap between "science" and "implementation".
- Direct connection to business results: the effect of each model can be calculated.
- Variety of tasks: from classic modeling to research ad-hocs and evaluation of new data sources.
- Opportunity to try new approaches and technologies on real data, not in isolation.
Hard Skills
- Python (confident command of functions, classes, OOP approach);
- Pandas, scikit-learn;
- SQL (indexes, window functions, CTE).
Soft Skills
- Flexibility in approaches to tasks;
- Desire to try new things and not be afraid of experiments;
- Ability to see new opportunities – in existing data or in state-of-the-art technologies – and apply them to current processes.
We Offer
- Participation in the development of an international fintech group.
- An international team and a dynamic work environment.
