After completing a graduate program at GE Power, I joined the GE Healthcare - Women Health - service team where I actively contributed to the digital transformation of the Service team at GE Health Care.
Where I Was
I was part of the service team for the mammography modality, based in Buc, working within an Agile environment.
What I Did
As the leader of the service team’s digitalization initiatives, I led various projects aimed at service digitalization. These projects included implementing predictive maintenance, enhancing databases, and creating automated tools to drive cost reduction initiatives. This role provided me with a comprehensive understanding of data processes, spanning from system-level considerations to data collection, pipelines, database management, modeling, and data analysis.
Predictive Maintenance
Predictive maintenance involves assessing equipment conditions through regular or continuous monitoring. The goal is to perform maintenance at the most cost-effective time before equipment performance degrades beyond a threshold. This approach minimizes unplanned downtime, benefiting both customers and suppliers:
- Customers experience reduced unplanned downtime, increasing machine utilization time.
- Suppliers cut costs by targeting specific issues, reducing unnecessary part replacements.
- Strong data quality is crucial for predictive maintenance. Models like linear regressions and neural networks can be developed to detect and prevent early failures with dedicated data infrastructure.
Databases Enhancement
Since predictive maintenance relies on high-quality data, it’s vital to develop data collection from sensors, databases, and infrastructure for parallel data collection and analysis alongside predictive maintenance model development. In my role, I contributed to developing new logs, databases, and monitoring techniques.
I worked with both SQL databases (MySQL, Postgres, Oracle) and NoSQL databases (MongoDB). I also created Python scripts for data collection and analysis, as well as used Spoon for pipelines and transformations. I managed data from end to end, from the system to the engineers using the data for project enhancement.
Automated Tools for Cost Calculations
With the influx of data and changing parameters, understanding which models can impact costs and service efficiency becomes challenging. To aid decision-makers in setting priorities, I developed automated tools for cost calculations. I built a Python tool based on Natural Language Processing (NLP) to classify and categorize costs, helping prioritize actions based on service data. Results were visualized using Sisense.
What I Learned
From a technical standpoint, I gained proficiency in Python and deepened my knowledge of databases. Working in an Agile environment also exposed me to new working methodologies.
Skills Acquired
- NLP
- Python
- SQL
- NoSQL
- Databases (MongoDB, MySQL, Postgres)
- Industrial Digitalization
- X-rays
- Mammography
- GE
- Predictive Maintenance