Position Senior DS To: Yandex Financial Services Gross Salary Range: 350-550 + bonus twice a year based on review results ~ 120% of salary Location: Moscow or St. Petersburg, hybrid, remote within Russia is possible

About us: ML in Fintech is all about machine learning for Yandex's financial products: from credit scoring and risk management to marketing models, recommendations, and GenAI assistants. The team doesn't just train models; they are responsible for their entire lifecycle — from data discovery and preparation, training, and experiments to deployment and runtime stability support. As a result, ML directly impacts the product, and the outcome is measured by business metrics, not just offline quality.

What you will do? • Customer LTV (Lifetime Value) Forecasting Using ML models, we predict the long-term value of each customer for the company. This allows us not just to segment the audience but to build a customer-centric strategy: optimize marketing costs, develop personalized offers, and consequently, directly influence key business financial indicators.

• Customer Churn Prediction Modeling Retaining an existing customer is cheaper than acquiring a new one. Churn prediction models are a tool for targeted and timely intervention, allowing us to not only reduce losses but also ensure a stable revenue stream from the core customer base.

• Building Uplift Models for Communication Optimization We determine which communication will be most effective for which users. As a result, we minimize spam annoyance and optimize advertising costs, increasing the effectiveness of each communication.

• Developing Personal Recommendation Models Personal recommendations significantly improve the user experience by helping customers find what they need faster and discover new possibilities within our services. This directly leads to increased engagement, product usage depth, and conversion into target actions, ultimately enhancing customer loyalty and revenue.

• Building User Transaction Embeddings We build neural network models that transform raw user transaction activity — payments in Yandex services, debit card operations — into vector representations of the customer's behavioral profile. These universal embeddings enhance the quality of models across a wide range of company tasks.

Requirements: • Understand the principles of classical ML and have worked with it • Understand the principles of DL and have worked with it • Proficient in Python and SQL • Have developed models and ideally deployed them to production • Ready to be responsible for production models in runtime

Will be a plus: • Use LLMs in day-to-day life or pet projects • Understand how LLMs work • Understand how RAG and agent-based systems work

Answers to 10 questions:

1. Data: YT - Yandex's Hadoop-stack analogue **2. Hardware: YT - 5 PiB, 5000 cores. Production: MLPlatforma - platform for deploying models to production. For work: Macbook
2. Scale of business impact: Responsible for all applied ML (except scoring) in Yandex Fintech: LTV, Churn, Communication Uplift Models, recommendations, balance models, etc.
3. Level of DS development in the company: Yandex - high level of DS development **5. Data Scientist role: end-to-end - responsible for the full development cycle: from idea challenge to model deployment to production
4. Manager's background: linkedin
5. How often will you be interrupted at work: 3 team meetings per week + 1-2 meetings with stakeholders
6. Career growth: semi-annual reviews with salary and grade revision **9. Prod/Research: 90% production development
7. Service or leadership function? service function from the start

How to apply: Send your CV and questions here: tg recruiter contact