Graduate ML Engineering Intern

EnerSys Delaware Inc.

1d • Onsite

About The Position

EnerSys is a global leader in stored energy solutions for industrial applications. The company has over thirty manufacturing and assembly plants worldwide, servicing over 10,000 customers in more than 100 countries. Worldwide headquarters are located in Reading, PA, USA, with regional headquarters in Europe and Asia. EnerSys complements its extensive line of Motive Power and Energy Systems with a full range of integrated services and systems, leveraging over 100 years of battery experience. This internship focuses on the development and validation of ML models for anomaly detection, predictive maintenance, and battery health estimation, with a particular emphasis on Remaining Useful Life (RUL) prediction for BESS and battery products. The intern will work with real industrial time-series and telemetry datasets, apply and evaluate a range of modeling approaches, and contribute to the data infrastructure that supports ongoing research and deployment. Engagement with adjacent projects in energy management and digital twin simulation is expected to be based on project priorities.

Requirements

• Currently enrolled in a Master's or PhD program in Computer Science, Electrical Engineering, Data Science, Applied Mathematics, or a related field — or recently graduated from such a program.
• Strong theoretical and applied understanding of supervised and unsupervised machine learning — including classification, regression, anomaly detection, and clustering methods.
• Proficiency in Python and core ML libraries (scikit-learn, PyTorch, or TensorFlow).
• Demonstrated experience with time-series analysis and modeling — feature engineering for temporal data, handling non-stationarity, temporal cross-validation, and sequence modeling.
• Solid data engineering skills — proficiency with pandas, NumPy, and SQL; experience managing and transforming large, complex datasets.
• Rigorous approach to model evaluation — including appropriate metrics selection, handling class imbalance, avoiding data leakage, and assessing overfitting in time-series contexts.

Nice To Haves

• Experience with anomaly detection or predictive maintenance — including statistical, ML-based, or hybrid approaches (e.g., Isolation Forest, autoencoders, or probabilistic methods); demonstrated application to real-world sensor or telemetry data is a plus.
• Familiarity with deep learning for time-series tasks, such as sequence modeling, attention mechanisms, or fine-tuning of pre-trained temporal models (e.g., Chronos, TimesFM, or similar foundation models).
• Interest in or prior exposure to battery systems, electrochemical degradation, power electronics, or equivalent experience with complex physical systems producing operational data.
• Familiarity with cloud-based ML platforms, particularly Microsoft Azure (e.g., Azure ML or Azure AI Foundry), or equivalent experience with experiment tracking and model management tools (e.g., MLflow, Weights & Biases).
• Exposure to optimization methods (e.g., linear programming, MPC) or physics-informed modeling approaches is a plus.

Responsibilities

• Develop and benchmark ML models for anomaly detection applied to battery and energy storage system telemetry, evaluating both statistical and learning-based approaches across varying data regimes.
• Build and validate predictive maintenance pipelines — spanning feature engineering, model selection, calibration, uncertainty quantification, and performance monitoring.
• Research and implement approaches to Battery Remaining Useful Life (RUL) estimation, including data-driven, physics-informed, and hybrid methodologies; evaluate generalization across battery chemistries, operating conditions, and degradation patterns.
• Evaluate deep learning architectures — LSTMs, Transformers, and CNNs — as well as time-series foundation models (including Chronos-2, TimesFM, and Lag-Llama) for forecasting and health estimation tasks; conduct systematic comparisons across accuracy, inference efficiency, and data requirements.
• Contribute to data pipeline development and feature engineering workflows, with attention to temporal alignment, missing data handling, label quality, and reproducibility.
• Support the digital twin platform and EMS framework through simulation contributions, forecasting model integration, and asset optimization modeling as project capacity allows.
• Engage in the full engineering rigor expected of production AI/ML systems — including unit and integration testing, model verification and validation, experiment evaluation, simulation workflows, data and output visualization, and technical documentation and reporting — as continuous activities throughout the project lifecycle.

Benefits

• Paid time off plus paid holidays
• Medical/dental/vision insurance plan
• Life insurance, short/long term disability, tuition reimbursement, flex spending, and employee stock purchase plan
• 401K plan

What This Job Offers

• Career Level: Intern
• Number of Employees: 5,001-10,000 employees