From Intel to Google
From Intel to Google, from hardware engineering to software engineering, many aspects of my work have changed. Some of these changes I still haven't fully adjusted to. Below, I've listed a few of the biggest differences. This might be helpful for those considering working in the Bay Area or making the switch from hardware to software.
Intel has a lot of work-related jargon. For example, in our memory department, some common ones I remember are: BKM (Best Known Method), POR (Plan of Record) which means something that has been agreed upon and recorded, DPM (Defect per Million) which means defect density, ECC (Error Correcting Code) used in algorithms that can repair errors, ESD (Electrostatic Discharge) which is static electricity that can damage circuits, LBA (Logical Block Addressing) which is a memory address, MLC (Multi-level cell) which is a memory unit that can store more information, and so on. Since I joined the company, I've kept a list of abbreviations. Every time I discovered a new one I would add to it. By the time I left, that list had at least 100 abbreviations. In addition to these jargons, there are also many product codes and experimental process codes, making the entire slide deck very hard to understand, very unfriendly to junior engineers.
There are also a lot of files, documents, and data in Intel. Whenever there was new product spec, important meeting materials, or factory testing reports, I would take screenshots of the new information and paste it into my Microsoft OneNote notebook, where I categorized it for easy searching later. Oftentimes, when the team wanted to find something, they would come and ask me. This notebook, which was several gigabytes in size, became an important asset in the team. One time, when my computer was replaced, this notebook was lost in the cloud, and my colleagues immediately suggested escalating the issue and reporting it to ensure the data was recovered.
Unlike meeting notes and documents, the experimental data is usually stored on the server's hard drive. The hard drive space was limited, and it would be full every now and then. When the data was about to reach the limit, everyone would receive an email from IT, imploring everyone to delete the unnecessary data; the names of the culprits who stored the most files would be publicly announced. Interestingly enough, the data at that time was almost all Excel spreadsheets, compressed in zip files. Those who needed the data would copy it themselves, decompress it, and then create charts in the spreadsheet. Later, the higher-ups decided that we should move the data to the cloud, have a unified interface to browse the data, and maybe even use AI to help everyone visualize the data in the future. Initially everyone thought it was a good idea: we no longer need to use Excel to draw diagrams, copy and paste to share with people, even thought that was how we did things for all these years. Then, after the data was migrated to the cloud, a group of senior engineers who had worked at Intel for a very long time realized that they had to re-learn the website interface to search for data, and decided they don’t like it. In addition, many of Excel's more advanced functions were not available on the website. Essentially, this migration failed, and started pointing fingers as which manager had the greatest responsibility.
What about the data that is already in the cloud? Because I was the only one in the team who could code, the manager gave me a task to write a "zip download tool", an exe file. When someone clicks on this exe, enters the desired experimental data, my tool would grab the data from the cloud and generate a zip file. So after going around a big circle, zip files are still everyone's favorite, and so was Excel.
After joining Google, I could no longer use Microsoft Excel or OneNote, etc. I was wondering how everyone took notes? Later, I realized that Google is indeed Google. You can find information by searching on Moma (internal Google), and you don’t need to copy the files in your computer. Everyone’s slides and documents can be found if you have permission. All the data is in the cloud. No one uses spreadsheets to create charts; we mostly use dashboards. Even numbers like coding time, seniority, how many lines of codes submitted of employees can be found in the cloud. The only thing that cannot be searched is confidential information that will affect stock prices, such as the number of YouTube users and revenue. Therefore, compared with Intel, the workflow is vastly different in Google.