Description

We are looking for an intern for a research project at the intersection of Machine Learning, representation learning, and chess data analysis.

The project is dedicated to developing a universal chess position embedder – a model that creates a compact and semantically meaningful representation of a position and can be used in various downstream tasks: from assessing position complexity to cheat detection and interpretable analysis of player mistakes.

Responsibilities

• Preparation and processing of chess game datasets
• Development and training of a contrastive embedding model
• Building a training pipeline
• Comparison with baselines (Allie, Maia2, behavioral stylometry)
• Conducting downstream experiments
• Interpretability experiments

Requirements

• Strong knowledge of Python
• Experience with PyTorch / TensorFlow
• Understanding of contrastive learning and representation learning
• Basic understanding of ML experiments (metrics, validation, baselines)
• Interest in research and academic work

Will be a plus:

• experience working with chess engines (e.g., Stockfish)
• understanding of the ELO system
• experience working with large datasets
• interest in explainable AI.

Conditions

• Comfortable modern office near Kutuzovskaya metro station
• Hybrid work format
• Corporate gym and recreation areas
• More than 400 educational programs from SberUniversity for professional and career development
• Onboarding program and manager assistance at the start
• Referral bonus for recommending friends to the Sber team.