Predictive Maintenance Engineer

Description:

We are looking for a pragmatic, production-minded Predictive Maintenance Engineer to build machine learning solutions that predict failures, estimate state-of-health (SoH) and remaining useful life (RUL) for UPS and battery systems. You will turn telemetry into reliable diagnostics and actionable alerts — owning the full ML lifecycle from data ingestion and feature engineering to model deployment (cloud and/or edge), monitoring, and continuous improvement. This role sits at the intersection of data science, MLOps and embedded/IoT systems and requires close collaboration with firmware, field service, product and cloud teams.

## Responsibilities
- Ingest, clean and align large telemetry datasets from UPS, inverters and battery management systems (voltages, currents, temperatures, event logs, charge cycles, alarms).
- Design and implement time-series feature engineering, anomaly detection, forecasting and RUL estimation pipelines.
- Develop and validate ML models (supervised, unsupervised, hybrid) for SoH, fault prediction and early-warning scoring.
- Build automated training, validation and deployment pipelines (CI/CD for models), and integrate models into cloud/edge inference environments.
- Work with firmware/edge teams to enable on-device inference (quantization, model size/perf tradeoffs) when required.
- Define evaluation metrics, acceptance criteria and A/B test plans; maintain model performance monitoring and drift detection.
- Produce dashboards, technical reports and runbooks for operations and field teams; translate model outputs into clear operational actions.
- Support root-cause analysis for field incidents and iterate models based on feedback and new labeled events.
- Establish data quality checks, instrument new telemetry where needed, and collaborate with hardware/sensor teams on meaningful signals.

## Qualifications
**Required**
- BSc/MSc in Computer Science, Electrical Engineering, Data Science or related field (PhD a plus).
- 3+ years of experience applying ML to time-series / telemetry data in production environments.
- Strong Python skills and familiarity with ML libraries (scikit-learn, PyTorch, TensorFlow) and time-series toolkits.
- Experience with data engineering / pipelines (SQL, Pandas, Spark, Kafka, Airflow or similar).
- Practical MLOps experience: containerization (Docker), CI/CD, model versioning and monitoring frameworks.
- Familiarity with cloud platforms (AWS/GCP/Azure) for model training and deployment.
- Excellent problem-solving and communication skills; able to explain model behaviour to non-ML stakeholders.