Data Scientist – Predictive Maintenance

This is a fully remote position, open to applicants in Brazil.

📋 Description

• Create time-series machine learning models for tasks such as anomaly detection, forecasting, behavioral modeling, and condition monitoring utilizing extensive industrial sensor data.

• Convert experimental models into dependable, maintainable, and scalable production systems.

• Develop metrics, diagnostics, and monitoring tools to ensure data quality, feature distributions, anomaly scores, prediction stability, model drift, and overall performance.

• Explore modeling approaches that balance predictive accuracy, generalization, adaptability, interpretability, and maintenance costs.

• Implement feature extraction, signal processing, and representation learning methodologies for industrial sensor data.

• Formulate hypotheses, design controlled experiments, benchmark various approaches, and assess models using temporal validation and realistic deployment scenarios.

• Enhance data processing, feature computation, model serving, and inference workflows for applications requiring high throughput and low latency.

• Assess cutting-edge methods and adapt promising techniques into production-ready implementations.

• Collaborate with software engineers, product teams, and lab specialists to validate models, troubleshoot failures, and enhance predictive maintenance products.

⛳️ Requirements

• Strong foundation in machine learning, with hands-on experience in developing models for time-series data, anomaly detection, forecasting, or similar sequential modeling challenges.

• Proficiency in Python and familiarity with essential ML and data processing libraries, such as NumPy, pandas, scikit-learn, PyTorch, or comparable frameworks.

• Experience transitioning machine learning models from experimental phases to production stages, including code organization, testing, performance optimization, and integration with data or inference pipelines.

• Comprehensive understanding of time-series principles, encompassing temporal dependencies, non-stationarity, seasonality, noise, and suitable validation strategies.

• Awareness of data drift and model degradation in production environments, including methods for monitoring distribution shifts, assessing model stability, and recognizing when model updates are necessary.

• Strong software engineering capabilities, including the ability to write modular, testable, and maintainable code while designing solutions that function reliably at scale.

• Experience in designing experiments and benchmarking models using relevant metrics, baselines, and validation approaches.

• Ability to analyze intricate, noisy, high-dimensional data and convert empirical insights into sound technical decisions.

• Proficient in both English and Portuguese.

🏝️ Benefits

• Competitive Compensation

• 30 days of paid annual leave

• Education and courses stipend

• Earn a trip anywhere in the world every 4 years

• R$1.035/month for meals allowance

• Health plan with national coverage and no co-payment required

• Dental Insurance: assistance with dental treatment for improved quality of life.

• Wellhub Membership: Access to a diverse array of gyms and training programs.