Burnout is a Neverending Struggle

Burnout is a funny phenomena which I hope that you, dear reader, never have to experience. However, it is important to understand it and know how to handle it. At this point I would say that about ⅓ of my professional career has been marred by burnout. I have gotten better at handling and resolving the issues causing burnout, but I wanted to share a bit of my experience with it.

The First is the Worst ☹️

My first time experiencing burnout was with my longest standing job at Epic. In almost every way, this has been my best and worst job. It was good in that initially I was learning a lot about delivering software products and good practices for managing and delivering code changes. At the same time, the practices in the expectations for delivery were on par with bad factory practices.

After all, “Epic is a software factory”. I was also 100% buying into the Kool-Aid message that we are “heroes helping heroes” and we should “paint the world red” (red is the color of the Epic logo). I will validate that building electronic medical record technology that is safe and usable is a difficult task that if done incorrectly will result in patient deaths. Epic does deserve credit for it’s focus on patient safety and user experience.

With that said, Epic doesn’t have a great focus on employee needs. One of the problems is the tendency to promote well performing engineers into leadership. This means that leaders are very technically savvy, but may not have all the well developed soft-skills necessary for the job. In other companies, they would separate leaders into a technical lead and a managerial lead. There was no such distinction at Epic.

All of that background is background information to understand how I burned out. I was actually promoted to be a small team lead (1-3 direct reports) for the note dictation functionality. This was originally shared by two teams, but was made into a small team for me to manage. I was actually fairly successful in managing the functionality and improving the stability of dictation. It was then that the dictation team was folded into the physician notes team.

And that’s when things started to get bad. My team lead was overloaded and unable to meet with all of his direct reports and manage the section (team of teams) he was also tasked to manage. So I unofficially had to have and additional 5 people as direct reports. Part of the issue was that I didn’t know how to say no to him and I was still trying to help. This massive increase in workload was causing me to work 55 or more hours. Time with the family suffered and the stress was building up.

It was in the middle of this that I had the brilliant idea to participate in an “Epixperiment”; a two week time to experiment with new technology and present any findings from it. At the time, we had somebody come through and present on a natural language model which could be used for interpreting commands. Think a combination of speech-to-text along with ChatGPT level of context awareness. Example: “Hey Epic, order Acebutolol 400 milligram capsules” would search and find Acebutolol medication for 400 mg capsules.

I was working with a colleague to work with an open source model, but found it too difficult to work with as many of the parameters are exposed for a research context. We instead found a small keyword detection program/code and integrated that with the test environment to implement a few choice commands. We had it tuned to work for a demo of what it could be used for, but on the stage the detection bugged out. Also during this time, I had to work on a performance improvement plan for one of my not-direct reports. I was a hot mess after that day.

Interestingly, I was offered the opportunity to lead Epic’s team in their language model integration with Dragon a proprietary model which was also experimented on. It was at that meeting where I expressed that I wanted to step down as a team lead. I was then shucked back to the note taking team as just a developer.

However, the fun was only just beginning. The same lead TL had made a promise to a customer for a piece of note taking functionality which had not been successfully delivered on despite being worked on multiple times in the past. Already in that year there was one developer who had tried to deliver it and chose to transfer teams instead of continue working on this project. So I was chosen to take on the task. It didn’t help that timelines were tight and there were a lot of safety and performance concerns with this functionality.

In the end, I created a design which was simplified the problem slightly to alleviate most of the safety concerns. I poured pretty much all of my effort into this project. I was burning the candle at both ends for nearly an 8 month period. I delivered the project, but I was a bitter man; cursing more, skipping common pleasantries, and just being a really negative person.