Proving Invisible Problems with Data
What do you currently do at AIO?
I am responsible for quality improvement and product engineering related to the development of uSD, eMMC, and UFD Controllers within the MPE Team. My role is to oversee and verify the entire process from chip design through packaging and final product reliability evaluation, working between the design house and the customers who ultimately use the products.
More specifically, I set up and analyze EDS (Electrical Die Sorting) at the wafer level, design boards and establish test setups for package reliability evaluations such as HTOL, ELFR, and LTOL, and perform yield analysis and abnormality (ABN) management based on CP (Chip Probing) test data.
At the package design stage, I also review NAND stack configurations and conduct SI (Signal Integrity) and PI (Power Integrity) simulation verification for USB and eMMC interfaces. Based on IBIS models and trace impedance conditions, I work with external simulation specialists to verify that board and package designs meet signal quality and power stability requirements under actual operating conditions.
What skills or capabilities do you consider most important in your role?
I believe the most important capability is the persistence to investigate an issue thoroughly and follow the evidence until its root cause is identified.
Most issues, such as abnormal wafer yields or abnormal circuit currents, do not reveal their causes immediately. A key part of the job is therefore to identify the root cause by systematically eliminating variables one by one, including WAT data and probe card resistance.
Another important capability is communication—the ability to bring together stakeholders with different perspectives around a common set of facts. Design houses often view an issue from a circuit design perspective, customers from a quality and reliability perspective, and our internal teams from a mass-production and yield perspective.
In these situations, rather than deciding whose view is correct, it is important to use measured data as a common basis for discussion so that all three groups can evaluate the issue based on the same facts. I also develop and use tools for organizing and visualizing data, which I believe is an important way to support this communication process.
What has been your most rewarding experience?
One of my most memorable experiences was receiving confirmation that we had passed all of the HTOL, ELFR, LTOL, ESD, and Latch-up tests that we had spent a long time preparing for as part of a recent reliability evaluation.
It was a project in which I was directly involved in multiple stages, from board design and test-condition setup to collaboration with external testing organizations. It was particularly meaningful because the results directly determined whether the product could be supplied to the customer.
I also find it highly rewarding whenever we identify the root causes of various failure samples one by one and see those findings lead to actual yield improvements and the prevention of recurrence. Proving an invisible problem through data is one of the most rewarding aspects of this role.
What would you like to achieve at AIO?
I would like to steadily increase the number of products, such as our eMMC Controllers currently in mass production, that earn customer recognition for their stable quality.
In particular, I hope to contribute to building team-level expertise and know-how for consistently analyzing issues that arise from process variations, including subtle abnormal current behavior whose causes have not yet been clearly identified.
I also plan to continue automating repetitive analysis and verification tasks using AI-based tools. We currently develop and use a range of tools in-house, including WAT data analysis tools, Yield management programs, Bin statistics management programs, Schematic and Layout verification automation tools, and database management tools for board and package deliverables. We are gradually shifting from a process in which engineers repeatedly review and interpret data manually to one in which tools identify signs of abnormalities automatically.
In the long term, I would like to further advance these automation tools and create an environment where engineers can focus more on root cause analysis and decision-making rather than repetitive tasks. Ultimately, my goal is to systematize the entire workflow—from initial verification at the wafer stage through package reliability evaluation and customer support—so that risks arising during new product development can be managed more quickly and accurately.
What advice would you give to those preparing to join AIO?
Semiconductor development involves far more effort in tracking down invisible problems and identifying their causes than in producing immediately visible results. If you are someone who notices an anomaly in a single line of data or a single point on a wafer map and asks “Why?” rather than simply moving on, I believe this role would be a good fit for you.
MPE work is also highly collaborative. It requires continuous cooperation with design houses, customers, and various internal teams. For this reason, the ability to explain technical issues in a way others can understand and coordinate different perspectives is just as important as technical expertise.
If you are not afraid of unfamiliar problems and are willing to collaborate with colleagues based on evidence, I believe you can grow significantly at AIO.