Data Science Lead Offline

About the Project
The company develops a data-driven analytics platform for utility providers, focused on improving water quality and operational efficiency. The platform ingests large volumes of IoT sensor data from field devices, combines it with laboratory and environmental data, and applies physics-informed and statistical models to identify anomalies, predict risks, and generate actionable operational recommendations.
The system models complex physical processes (e.g., chemical dosing optimization, maintenance prioritization) and aims to deliver enterprise-grade B2B software, integrating domain knowledge from water scientists with scalable AI solutions.

We are looking for a Data Science Lead to drive the modeling and inference layer of the platform. You will design, validate, and implement advanced physics-informed, stochastic, and causal models using time-series and sensor data. The role requires strong statistical background, hands-on Python experience, and an ability to translate complex scientific models into production-ready machine learning components.
You’ll collaborate closely with the CTO, COO, and domain experts, working side by side with the hardware (IoT) and software engineering teams to ensure models are robust, explainable, and production-ready.

Requirements:
- 5+ years of experience in Data Science or Applied Statistics;
- Strong expertise in Bayesian methods, causal inference, stochastic modeling, and process/physics simulation;
- Hands-on Python skills (NumPy, Pandas, Scikit-learn, PyTorch/TensorFlow);
- Experience with production ML practices — pipelines, model versioning, monitoring, and validation;
- Proven work with sensor data or industrial systems (IoT, environmental, or manufacturing);
- Excellent communication and ability to collaborate with scientists and engineers;
- English — Upper-Intermediate or higher.

Responsibilities:
- Design and validate statistical, stochastic, and physics-based models for water quality and reliability analytics;
- Build and optimize time-series feature pipelines, ensuring robust preprocessing, data validation, and outlier detection;
- Integrate models into the platform’s production environment with proper testing, monitoring, and retraining cycles;
- Collaborate with data engineers and domain experts to interpret laboratory and field data for model improvement;
- Ensure model explainability, uncertainty quantification, and interpretability;
- Contribute to digital twin and simulation model development combining real-time sensing and process knowledge.

Working conditions:
- Paid vacation (15 business days per year);
- Flexible working hours (8 hours a day);
- Paid workshops, conferences, courses.