Vertex AI/ ML Engineer

Description:

Client is seeking an ML/AI Engineer to develop, fine-tune, and operationalize machine learning models that predict track nonconformities on the subway system. Working within a Google Cloud Vertex AI environment, you will leverage sensor data (vibration, audio, location) captured by Pixel phone hardware kits on revenue cars and ground-truth defect records from MTA''''s Hexagon system to build predictive models supporting the TrackInspect application.

Key Responsibilities

·         Perform feature engineering on sensor and operational data to identify patterns correlated with documented track defects in Hexagon (HxGN)

·         Leverage and refactor existing prototype ML models for pilot-scale use; validate performance and suitability for production

·         Utilize Vertex AI Workbench to train, retrain, and version ML models; track experiments and metrics

·         Conduct hyperparameter tuning and model validation activities to optimize nonconformity prediction accuracy

·         Deploy trained models for daily batch inference; connect deployed models to batched sensor input pipelines

·         Implement model monitoring, alerting, and a continuous feedback loop from Track Inspector inputs captured in HxGN

·         Build and maintain CI/CD pipelines and multi-environment deployment strategy for current models

·         Build the BigQuery ''''Predictions – User Friendly View'''' incorporating location attributes, prediction metadata, and feedback fields

·         Assess model results against known track defects within HxGN and present findings at model evaluation meetings with MTA

·         Proficiency in Python (scikit-learn, TensorFlow or PyTorch) and SQL

·         Hands-on Vertex AI experience (Workbench, training pipelines, model deployment)

·         Strong understanding of feature engineering, model validation, and hyperparameter optimization

·         Experience with time-series or sensor data (vibration, audio, or accelerometer signals a plus)

·         Familiarity with BigQuery as a data platform for ML pipelines

·         Experience with geospatial ML or GPS/LRS-based positioning models

·         Prior work in predictive maintenance, anomaly detection, or IoT sensor analytics

·         Google Professional ML Engineer certification

·         Understanding of transit or infrastructure asset management systems

·         Experience with Looker Studio or similar BI tools for ML output visualization