On the role of CTO
The ins and outs of being a CTO at Reddit
On being a CTO at Reddit:
I think the CTO role at Reddit is primarily kind of a chief architectural role. I have an OCTO (Office of the CTO) group and the tagline for that group I usually refer to it as extreme tactics and extreme strategy. So, either long-term planning or drop in and it’s like fix a thing and move on which is a lot of fun.
When a team scales:
You can keep track of the individual stories of about 50 people and you can keep track of about 150 people’s names. So, if you get past that point, all of a sudden you will have to rely on a trust network to make things work.
One of the nice and scary things about being CTO is I always have to make sure to hire people that are better at their job than I am. You have to rely on people’s ability to make decisions so I prefer to provide guidelines and let people make the decision.
Scaling through values
Back in the early days, Reddit has seen exponential growth, doubling size every three months, which forced the whole team to learn continuously about the tech, live. This enabled the engineers to instill a culture of learning up to now.
There are benefits and detriments to diluting with new talent. You get the advantage of new people who are bright-eyed and bushy-tailed. But at the same time, you have to make sure to take the time to — it sounds severe — indoctrinate people with the company culture.
How is Reddit building those values?
We usually do it at the executive level. I think a lot of the values end up being the results of Steve Huffman, our CEO, mulling on it for a while.
We generally have a brainstorming session where we discuss what we should do to tweak the values. Part of the values are quite easy to change as we already iterated on them many times. Then, there are values we keep, not because we need to work on them, but because it’s essential for Reddit that everyone thinks about them at all times.
How do these values include a body of knowledge and feedback brought by executives?
For us, it’s been ultimately about transparency. When we have large discussions, we try to get it all documented, get everyone to write their feedback back into the document, so they can read it too. Documenting has been critical to making sure that that feedback is heard and acted on. That works really well to make sure that your message is getting across and I think it also encourages the person who’s delivering it to wordsmith it and think about what they are trying to say.
Scaling the organization
There is a set of engineers who are sufficiently competent that they don’t need to write unit tests, but that is a vanishing small number of people. If anything, the utility of unit tests as a means for us to make sure people do not repeat mistakes.
In order to help people level up, Chris shared his process to find the right playbook:
I think for leveling up, it’s been a group effort across our leadership team and our management team. It really comes from the idea that you can do it the hard way and crank at the wheel, or you can find people who’ve done it before and learn from them. In this case, it’s been definitely leaning on people who’ve done it before and building those playbooks from that.
One-on-ones are valuable only if they are well-structured
IWe need to make sure that the one-on-one has a structure and that there is a path towards having the career advancement discussion on a regular basis. Not everyone thinks about their career or at least doesn’t say they are and so if you don’t have that outlet to actually talk about your career it can be very stressful to the actual IC.
His focus is more about spotting a capacity to reflect and analyse, instead of degrees.
Besides tech, personal skills are an essential quality he seeks in new hires:
We also definitely look for humility as a major component. I have worked with the self attributed rockstar engineers before and the hard part there is oftentimes if they come to us as a rockstar, they tend to be demotivating.
Chris encourages interviewers to give the same set of questions over and over again. Doing so will help them improve their interviewing skills:
You have to be able to see the pitfalls people fall into when they’re trying to solve the problem and know when you have to give a bread crumb to help them get a little further.
Giving breadcrumbs has another advantage:
I find breadcrumbs and whether or not the interviewee accepts them to be one of the most valuable signals you can get. It shows if the person listens to your feedback and if they are reasonable. I