Skip to content

Career thoughts

This document should be considered a living document that will gradually grow and change over time. Parts of this were originally pulled from the drunk and sober Reddit posts and edited. Other parts are from more recent communications and thoughts. Eventually I will want to reflect some level of what I want my career progression to be in a more directed sense.

Health

  • Nothing can replace your mental health and happiness. Treat those as the most important, and if those choices aren't respected, you're in the wrong place.
  • Be kind, both to others and to yourself. You are your harshest critic, but don't let that get in the way of your successes. Be proud of what you've accomplished.

Staff engineering

Senior technical roles aren't described as much, but there are some resources that capture that information, including StaffEng. There is a book as well, which I may buy after reading some more on the website first. Some general thoughts are captured below:

  • A staff engineer could be "wide" and have support across a bunch of various subteams across the organization, or "deep" and focus on one or two areas. Examples of "wide" could be product design, while "deep" is more likely to focus on security. I would probably enjoy the latter, since it would help with focus and expertise.
  • Staff engineers may have a few direct reports, but for the most part should not have many (or any).
  • Finding energizing work will keep me in a job for a long time, where I am pursuing impactful work. Bronwyn has also mentioned this for me to think about and follow.
  • Impact might be measured more by who is doing the work that you're supporting.
  • I think a tech lead or architect (see StaffEng) sounds interesting. I have the light experience of a technical lead from White Ops, which was both good and bad. At the time, since I had to be both a tech lead and a data scientist, that was one of the major underlying issues that led me to not enjoy that work as much.
  • Take the time to understand the status quo before trying to make changes. Be diligent and deliberate, and you will get results.
  • Deliberately accumulate expertise by doing valuable work. Focus on work that matters, do projects that develop you, and steer towards companies that value genuine experience.

Job hunting

During my interview process at Duo, I asked most of the people I interviewed with, “what would make you want to leave Duo?” The answers I got were varied and honest, and I’d like to take the time to write down what would be those triggering events for me at Duo. Much of this will be influenced by my previous roles.

  • Leadership that doesn’t respect or follow the company values
  • Leadership that doesn’t reward research and development into new approaches, but instead focuses on “gamified” KPIs to a detrimental extent
  • Signal collection apparatus used to support tracking, targeting, or other privacy-breaking activities on the internet
  • Teammates that don’t support or help each other, or an overall feeling on competition or aggression between employees
  • Lack of trust in the managerial relationship
  • Decreasing acknowledgement or care about diversity and inclusion at the company
  • Lack of humility in people following a mistake, or lack of psychological safety to be humble following a mistake
  • Job function purely serves sales or marketing versus improving the product
  • Lack of career improvement or growth potential

The above is subject to change over time, and not any single one of those on its own would be enough to get me to leave. Likely, if things turn for the worse like they did at White Ops/HUMAN, multiple points above will expose themselves around the same time, or those that were always present will become more pronounced.

Additionally, I still should consider a light job search after a few years’ time to ensure that I am still learning and growing in my career. At some point, I will need to consider what the next steps for me are: manager or principal IC. I’ve been developing in both, and could potentially step into either if necessary. Managerial work is a different skill set, and I would need to work to improve those skills in a directed fashion.

  • If you're unsatisfied at your job, that's a sign you should start looking elsewhere.
  • If people are assigning blame instead of figuring out the problem then fixing it, that's a sign you should look elsewhere.

Productivity

  • The amount of time worked has almost no bearing on the quality of the work. Most weeks, don't work close to 40 hours unless you feel like you are doing high-quality work. Take it easy on yourself.

Python and coding

  • Good code is simple, readable, and tested code. A student or new employee should be able to understand it. Better code is less code. The best code is no code at all. Find ways to reduce the amount of code you and everyone else writes.
  • Use mypy for performance improvements if you don't already have highly-optimized code. For high computational code, use numpy and CUDA-supported frameworks.
  • If you have a production pipeline, avoid pandas due to the high overhead in performance. Anything in this should either by in SQL or some scalable non-Python system if it needs a lot of transforms and such. Ideally, your "production" Python code will be purely model serving, or an offline framework that's been developed.