About the role
We’re looking for a Deep Learning Researcher to help build and ship production ML/DL models for time-series analysis and forecasting of financial-market behavior. We’re interested in people who have already built and deployed time-series forecasting models in production and want to apply that experience to quantitative finance, crypto markets, and algorithmic trading. Trading experience isn’t required — what matters is genuine interest in financial markets and a desire to go deep.
What you’ll do
- Build and ship production ML/DL models for time-series analysis and forecasting
- Research and apply modern Deep Learning architectures for forecasting and sequential modeling
- Analyze large volumes of historical and streaming data
- Experiment with new architectures, features, and training methods
- Optimize model quality and robustness in production
- Work with the Quant Research team to apply models to capital management and trading strategies
- Track research in Deep Learning, Time Series Forecasting, and AI
Requirements
- Degree from a strong technical university (MIPT, HSE FCS, MSU, ITMO, SPbU, Bauman MSTU, Innopolis, or similar)
- 2+ years of commercial experience as an ML Engineer / Deep Learning Engineer / ML Researcher / Data Scientist
- Experience building and shipping production ML/DL models for time series
- Experience shipping production ML/DL at a large tech company (e.g. Yandex, Sber, T-Bank, VK, Ozon, Avito, or similar) is a strong plus
- Deep understanding of modern Machine Learning and Deep Learning methods
- Strong Python
- Hands-on with PyTorch, Scikit-Learn, CatBoost, LightGBM/XGBoost, and Optuna
- Experience with time series, feature engineering, and model evaluation
- Our team works in Russian; English at a level to read documentation and research papers
Nice to have
- Interest in financial markets, crypto, or algorithmic trading
- Personal trading or investing experience
- Experience with reinforcement learning
- Experience with market data or high-frequency time series
- Publications, Kaggle, competitions, or research projects in ML/DL