Academia.edu Career Growth & Development

What’s it like to work on the engineering team at your company?

I work on a lean team of five engineers building Academia.edu’s AI co-scientist. It’s pretty fast-paced. We launch new experiments almost every week, look at how people actually use them and adjust from there.

Because the product is still new, most of our useful learning comes from building something quickly and getting it in front of real users — not discussing it for hours in a meeting room. The team has a lot of autonomy. If someone has an idea, they can usually make a rough version, share it with the team and test whether it’s actually useful.

Our backend is a Ruby on Rails monolith, but the product now spans the web, browser extensions and desktop apps. AI coding tools have made it more practical for a small team like ours to work across that range. We’re less restricted by the technologies we already know, so we can choose what makes sense for the user instead of forcing everything into our existing stack.

The tradeoff is that not every experiment works. Actually, a lot of them don’t. But we try not to get too attached to something just because we spent time building it. The goal is to learn quickly and use that learning to decide what to do next.

How does your team support learning, upskilling or technical growth for engineers?

We have a high-challenge, high-support feedback culture. That depends a lot on trust. It means having a difficult conversation when something isn’t working but also making sure we have the support we need to improve.

Our learning happens through experimentation. Everyone is free to try different AI tools and workflows and we regularly share what worked and what didn’t. It’s informal and practical. Someone might show how they used a tool to understand an unfamiliar codebase, work through a product idea or build something outside their usual technical background.

Freedom to experiment matters. If something feels clunky, I don’t have to wait for permission to improve it. I can try something, see whether it helps and share what I learned. I use Claude as a thinking partner to talk through an idea, which helps me find the gaps.

I stay current through podcasts like “Lenny’s Podcast” and our internal AI Slack channels, where we share tools, articles and examples from across the industry. Working on our co-scientist feature gives our team a practical reason to stay current. Part of our job is figuring out which new ideas are genuinely useful to researchers and which ones are mostly hype.

What kinds of projects or challenges help engineers build new skills on your team?

Our team mainly comes from Ruby and web application development, but we’ve built a desktop application, a browser extension and parts of our product in Python. AI coding tools don’t remove the need to understand what we’re building; rather, they make it easier to get started in an unfamiliar area and learn by doing.

Engineers also get to own work across the full product moving between back-end code, a desktop client, infrastructure and the user experience in the same project. It can be uncomfortable because there isn’t always an expert nearby with the answer, but that’s where a lot of the learning happens.

We experiment a lot, but everything we build should eventually make contact with a real user. If the response isn’t positive, which happens more often than we’d like, we try not to defend the work because of the time we already spent on it. We look at what we learned, decide whether the idea needs another iteration and move on when it doesn’t.

That combination of broad ownership and direct user feedback has been the biggest source of growth for me. You’re learning a new technology and also learning when it’s useful, where it breaks down and whether it solves a problem anyone really has.