Rachel Gardner, 2018 Summer Intern, Stanford University Computer Science, Class of 2020
In looking for a job, there is a constant question: big company vs small company? Rather than answer this question, I chose “all of the above” and interned at a medium-sized company (Silicon Labs), a large company (NVIDIA) and now a small company (BabbleLabs), one after the other. The first and most obvious difference is that I always have to explain what BabbleLabs does, as “I work at a deep learning startup” usually invokes a fair amount of interest (along with a few knowing smiles). In case you were also wondering, Babblelabs is a speech processing company, using advanced neural networks for cloud and devices. It’s less than a year old, but has already launched its first product (5 months after raising the first $4M).
With such a small company (about 9 in BabbleLabs’ San Jose office), I was immediately treated as a full-time employee, with all that brings. In the often unstructured environment of a startup, I found that my past work at larger companies gave me the experience to impose my own structure: setting goals for the internship, calling meetings to discuss milestones, etc. It was clear that the decisions of the founding team were similarly influenced by experience with more established companies, both in terms of how to do things and in terms of what to avoid. Because of the small size of the company, I had the opportunity to learn directly from this experience. At lunch everyone sat at the same table, discussing the same topics (work-related and otherwise). I also had weekly meetings with the CEO to discuss the broader entrepreneurship questions facing the company. Where in the past I had hung out exclusively with interns, BabbleLabs gave me an opportunity to learn from people with a lifetime of industry experience.
In looking for an internship, it is easy to get caught up in the amenities: the ping pong tables, the climbing wall, the free food. However, I was surprised to find that I didn’t miss those as much as I had expected. Size “extra small” suited me just fine. Though I had anticipated that the lack of a subsidized cafeteria would feel like a noble sacrifice, I realized that I didn’t mind the company-wide field trip to the little Polish café down the street. Where I thought I would miss having a space to myself, I found it valuable to be right in the thick of things. I enjoyed being able to take off my headphones and listen to coworkers discuss product strategy, hiring, or marketing. Of course, I also found out that I should have bought noise-cancelling headphones for those times I didn’t necessarily want to hear about the logistics of launch 2.0.
It would have been a good investment, as most of my days were spent head down, headphones on. After the flurry of mailing lists characteristic of a large company, my inbox felt eerily silent. The lack of distraction was quite welcome, however, as it meant I could attack the mountain of tasks ahead. When I joined the company I was the first one working in my particular area. This meant I was also putting together the framework as I went. Where my friends were still reading coworker’s code, or jumping through hoops to get access to the right repos, I hit the ground running on day one. No need to wait for a company-issued computer, ID badge, or account. But I needed that headstart, because I had a whole lot of scaffolding to write from scratch. This was a good learning opportunity for me, but it wasn’t always quite the sexy machine learning one might imagine.
When I joined, I was given the broad goal of “doing something cool with accents.” In classic start-up form, I got to figure out what that meant and how to get there. I needed to put together my own dataset, which meant a lot of working with contractors and figuring out how to clean the data. In a way, I had to do the same type of goal optimization we do in deep learning: I had to very carefully set incentives in order to get what I was looking for, and sometimes I got output I didn’t expect! In parallel, I was reading every paper I could get my hands on (as well as the steady stream of links coming from the Poland office). I began training models with the handful of datasets I had, gradually building up my own PyTorch training framework as I went and adding in new data as it came in.
Even as I was building my small part, the company itself was growing. Maybe the most exciting part of joining a small company is watching it get bigger. It’s weird to think that just a few months ago we had our launch party announcing Clear Cloud, our first product. At the event, I had the opportunity to chat with investors, media, and potential customers. Our group picture from the event will go up in the brand new office, looking out over the San Jose skyline alongside our logo. While I won’t be around to work in the new space, it’s exciting that a part of my contribution will continue with the company.
With BabbleLabs as my last datapoint in this survey of company shapes and sizes, I realized that the people you work with and the projects you work on far outweigh the importance of any other company characteristics. A larger company may mean you have more options in terms of people and projects, but at a small company you know exactly what you’re getting. There are certainly tradeoffs, but at the end of the day it’s the people themselves who are most important. That’s why I know I will be coming back to visit, even as BabbleLabs grows out of being “extra small.”