The posting
We are partnering with a leading player in the data centre space in a search for a Senior Data Scientist, Predictive Maintenance to build and run machine learning models that predict failures and give early warning of degradation across data centre and industrial assets, using sensor data to forecast asset health and estimate time to failure.
Responsibilities:
- Build and validate models on operational sensor data to predict failures, track degradation of mechanical and electrical equipment, raise early warnings, estimate remaining useful life and forecast time to threshold.
- Apply and benchmark a range of techniques against one another, including anomaly detection, change-point detection, survival analysis, time-series forecasting and representation learning.
- Choose between modelling approaches on the basis of measurable performance criteria and documented experiment results, so that every selection rests on evidence rather than preference.
- Weigh detection performance against false-positive rates, explainability and the demands of production deployment.
- Devise strategies for environments with no failure events.
- Tackle limited labelled failures and class imbalance, drawing on weak supervision, proxy labels, transfer learning and failure-data sourcing.
- Set up validation methods that measure model effectiveness accurately even when failure examples are scarce.
- Develop hypotheses and design experiments that test competing approaches against a defined baseline, quantify improvements with objective metrics, and record findings and limitations.
- Carry models through deployment in both development and production environments.
- Track precision, recall, drift and data quality on an ongoing basis.
- Define the criteria for retraining, rolling back and retiring production models.
- Diagnose drops in model performance and put corrective measures in place.
- Partner with product managers, software and platform engineers, and data centre subject matter experts who bring mechanical and electrical engineering expertise.
- Frame operational issues as machine learning problems.
- Fold domain knowledge into how models are developed and validated and how alerts are interpreted.
Requirements:
- Bachelor’s or Master’s degree in Engineering, Computer Science, Data Science or a related discipline, with a specialisation in AI/ML preferred.
- 5-8 years of relevant experience in data science or applied machine learning, gained in a large software development organisation or an end-user engineering or industrial environment.
- At least 2-3 years of hands-on experience building condition monitoring, predictive maintenance, failure prediction or anomaly detection models on engineering, operational or IoT sensor data.
- Good working understanding of common machine learning approaches and the reasoning behind them, covering classification and regression models, time-series models, anomaly detection methods and model evaluation techniques.
- Technical depth and a strong research mindset, with well-developed experimentation skills, the ability to work where failure examples are limited, and rigour in validating models under real-world conditions.
- A mechanical engineering background, or strong exposure to mechanical or industrial systems, is an advantage.
Interested candidates may email your resume to:
Alycia Lim
EA License No.: 23S2032 | R24124179
We regret to inform that only shortlisted candidates will be notified.
