Data Scientist - AI/ML Predictive Maintenance Platform

Role Overview

We are looking for an experienced and practical Data Scientist to contribute to the development of an advanced AI/ML-driven predictive maintenance platform. This position involves working with both traditional and generative AI methods applied to real-time Internet of Things (IoT) sensor data collected from a variety of physical assets. The ideal candidate embraces fast-paced development cycles, takes complete ownership of experimental processes, and emphasizes delivering solutions beyond prototypes into live production environments.

Key Responsibilities

• Design, develop, train, and optimize machine learning models utilizing time-series sensor data aimed at identifying anomalies and forecasting equipment failures across multiple asset categories.
• Investigate and implement a range of methods including forecasting, anomaly detection, change point detection, survival analysis, and representation learning, selecting the most effective technique tailored to each use case.
• Apply hybrid modeling strategies, such as integrating statistical thresholds with machine learning models or combining changepoint detection with LSTM/Transformer-based predictors, where beneficial.
• Ensure rapid iteration by delivering functional machine learning models into development or production stages approximately every one to two weeks to facilitate continuous feedback and enhancement.
• Evaluate and optimize models by balancing accuracy, interpretability, computational resource requirements, and deployment limitations.
• Manage ongoing monitoring of model efficacy in operation, focusing on metrics like accuracy, drift, and recall.
• Develop and implement strategies for retraining and rolling back models to address challenges arising from data and model drift or unusual cases.
• Create dashboards and alert systems for monitoring real-time performance degradation of models.
• Proactively suggest retiring or updating models as necessary.
• Communicate clearly the functionality of models including threshold settings and decision rules.
• Share all results openly, including experiment outcomes, performance comparisons, and the rationale behind model choices.
• Collaborate effectively with cross-functional teams including product management, engineering, and stakeholders with non-technical backgrounds, ensuring consistent ML decision-making and communication.
• Maintain concise but comprehensive documentation focusing on critical code, decisions made, and modeling strategies following lean principles.

Additional Information

This role operates within the Connectivity business segment and specific operating division platforms.