In the writing your resume course, I realized it is best to keep the resume simple and non-technical for recruiters to view. I used to have a lot of bullet points like Reduced the runtime by 9x meaning I was able to reduce on average runtime of 45 seconds to 5 seconds. Looking back, I don't think this makes much sense. I also wrote how I improved the time complexity from O(n) to O(1) which I deleted now.
What is the best way to write these accomplishments? Is improving runtime even a good data point in a resume? Should I write it to something like improved the runtime of a program from 45 seconds to 5 seconds? Or does improved by 9x make sense here?
Aside from this runtime issue, I'm wondering what are some good metrics in a resume. I realize the number of users is really impressive from Alex's resume, but what are some other good metrics?
The best way is to tie it to business value at the end of the day. This is often pretty hard, but did reducing the latency improve some KPI? For example:
Amazon study: Every 100ms in Added Page Load Time Cost 1% in Revenue
Thank you that is a helpful insight. In a case where such metric is no longer available because I'm out of the team, would it make sense to include latency in the resume?
I also feel like it's hard to get my experiences to apply to such article since I don't work on the main retail website. Some of my projects include data quality issues and improving the runtime of an auditing process from like 3 hours to an hour. What are some "KPI"s that I could look into?
As Sai mentioned, try to connect everything you did with business impact. Latency improvements are often very valuable as end-users hate waiting around for things to load. If you improved the layout on the front-end or optimized a critical product back-end API to load faster, that's obvious impact as the user can now access the product and its features faster.
The tricky part is you're reducing latency for something far-removed from the end-user like some nightly maintenance cron job or something where it doesn't really matter if it finishes in 2 or 20 minutes.
...improving the runtime of an auditing process from like 3 hours to an hour
"Auditing" generally refers to something done by a human, so claiming a win here is pretty easy. Figure out how many audits happen per day/week/month/year, multiple that by 2 hour time-savings, and now you can say something like "...saving the company 1,000 engineering hours per month" (I'm assuming engineers are the ones doing the auditing here).