I am finally done with this round of recruiting. I have no more phone screens or interviews scheduled. I started looking externally three months ago, talked to about 25 companies, phone screened with about 15, had about 5 onsite interviews. I had an offer from one company and am waiting to hear back from two companies.
The process was quite stressful, especially when I failed an internal interview for a manager role at Netflix. But the numbers don’t look that bad, considering that I am not currently a manager. Hopefully, I will be able to get an offer I am excited about at the end.
Software Engineering ⚙️
Mind the platform execution gap
This article is a good primer for what it means to be a platform: trust is the most important asset when a team builds a platform for other teams to rely on. Building software is complex enough that those teams do not need another variable in their work.
The article focuses on developer productivity platforms, but much of it applies to product platforms as well.
Unit Tests give you Observability!
A code coverage metric is not important. It’s a poor proxy for how well we understand our application. Instead of mechanically covering all lines of code, we should focus on the risky parts of code.
Tech Lead Management roles are a trap.
I did “full-stack” management when I was a manager because the startups I worked for didn’t have pre-existing structures: no career ladder, no recruiting team, and no defined hiring process. So I focused on creating those processes and less on managing individual performances. I also had to spend some time coding, which took more time away from performance management.
So when I look back, I haven’t communicated expectations and provided feedback enough. I don’t believe a tech lead role is a bad way to start managing (contrary to the article), but one should be aware of the risk and intentional about learning how to manage properly.
The Human Side of Elixir
I am against choosing a cutting-edge technology (Elixir, for example) over a boring one (Java) just for the sake of it. But doing so selectively could help to attract engineers to your team. You can’t compete on compensation, so you find another reason to choose your team, such as cool tech. My previous team was able to hire a top-notch engineer because we worked with serverless functions before it became mainstream.
A set of visualizations showing how biased inputs into the model training lead to biased results, even if the protected information such as gender is kept away.
A good call out that we won’t have “the Metaverse” but will have many competing metaverses.