I came into my Data Analyst Engineering Internship at Toshiba America Electronic Components expecting to work alongside Greg Shumaker and Cheng-Ju Wu to improve the data pipeline for Hard Disk Drive (HDD) health logs so it could handle a new log type, and, hopefully have time for a few other data projects.
However, my expectations were surpassed when, due to a data pipeline project delay, I was given the opportunity to expand our team's machine learning efforts, build interactive dashboards to make our drive telemetry easy to explore, and help my coworkers automate a reporting process that previously took days.
I feel incredibly lucky to have had the chance to be at the forefront of a new data effort on our team, surrounded by a supportive, curious, and driven group of people. I was given a large amount of ownership, which helped me grow quickly. I knew the modeling techniques and data processing tools from my graduate studies, but getting to use them on a real, messy, and high-stakes problem taught me more than ever before.
Most of my summer was centered on three connected projects. The first was a machine learning model that predicts hard drive failures before they happen. By learning patterns that precede failure across millions of telemetry snapshots, the model can flag at-risk HDDs days or even weeks ahead of time. The second was a modern, interactive dashboard that lets our teams explore the health of a HDD fleet at a glance, from fleet-wide trends down to a single HDD's full history. The third was automating a monthly telemetry report that was previously assembled by hand, turning a multi-day task into a single command.
I had the chance to present these findings to multiple executives, from the San Jose office’s VP of Planning and VP of Engineering to leaders in Japan interested in the work we were doing. Sharing something I had built with people across the company, and seeing them engage with it, was one of the most rewarding parts of the summer.
There are a few people I want to thank for making this experience memorable and invaluable. Thank you, Greg, for giving the best crash course on hard disk drives to someone who had never previously worked in hardware. Within a few days, I was able to understand the base level of HDD mechanics, logging standards, and more. Since you were learning some of the machine learning techniques alongside me, the questions you asked throughout the process helped me hone my communication skills and catch inconsistencies I would not have otherwise caught. Thank you, Cheng-Ju, for being so generous with your time and knowledge. You were encouraging, always willing to explain the details behind the data, and never too busy to drop what you were doing to work through something with me. Thank you to my manager, Robert Yeo, for encouraging me to do great work, helping me hone my stakeholder communication skills to distill complex and unfamiliar topics into digestible bites, and giving me the opportunity to grow by sharing my work with people throughout the company. While these three were the people I worked with most directly, my biggest takeaway from working at Toshiba is that everyone is kind, driven, and will go out of their way to help you. From the interns to the executives, there is a real sense of community.
Toshiba's HDDs generate a large amount of data in the field, and this summer only scratched the surface of what that data can tell us, from predicting failures to understanding HDD lifespan, workload, and quality. I hope to have the opportunity to continue making an impact at Toshiba in the future through data processing, modeling, and reporting.
