Data Scientist
Sunlux Technovation · May 2025 – Present
- Collected and analyzed high-volume, live multivariate sensor data (temperature, pressure, flow rate) to monitor boiler system performance and detect operational anomalies across multiple subsystems.
- Engineered data cleaning and preprocessing pipelines using Python, Pandas, and NumPy, handling missing values, outlier treatment, and normalization to improve sensor data accuracy and consistency.
- Applied unsupervised anomaly detection techniques including Isolation Forest, DBSCAN, and Local Outlier Factor to identify fault patterns without labeled failure data.
- Created exploratory data analysis workflows and time-series visualizations using Matplotlib and Seaborn to communicate insights to engineering stakeholders.
- Developed automated monitoring scripts using rolling statistics and threshold-based alerting to promptly flag abnormal sensor readings and reduce manual monitoring effort.
- Integrated FMUs and OPC UA data acquisition pipelines, designed scalable ETL processes, leveraged Memcache for in-memory caching, and validated end-to-end data integrity across simulator and database layers.



