Data Scientist

Data Scientist at ROBOTMIA, Novosibirsk, hybrid. Requires 3+ years in DS/ML, ClickHouse, clustering, neural network architectures for classification, Python/PyTorch/TensorFlow, ClearML, alerting, Git/Docker/Kubernetes.

3+ years of DS/ML experience, focused on results rather than process Ability to work with large databases (ClickHouse) and write complex analytical queries Experience with data clustering (K-Means, DBSCAN, hierarchical methods) and tuning these algorithms Ability to select and configure neural network architectures for classification tasks (CNN, RNN, LSTM, Transformer, etc.) Experience solving classification tasks with strong class imbalance (sampling, weighted loss, synthetic data) Python, Pandas/NumPy, Scikit-learn, PyTorch/TensorFlow Audio libraries: Librosa, torchaudio Experience with ClearML (MLflow, Weights & Biases as alternatives, but ClearML is a plus) Experience developing automatic alerting systems (drift monitoring, anomalies, production metric drops with auto-reaction) Git, Docker, Kubernetes (understanding of CI/CD for ML) Nice to have: Experience with RNN, LSTM, TCN, Transformer, WaveNet architectures Solutions for Source Separation or noise reduction Participation in Kaggle with imbalanced classification tasks Experience with streaming data and real production systems

Hybrid work in Novosibirsk.