Anthropic Machine Learning Engineer interviews, decoded.

The interview process was pretty structured and overall a good experience. I started with a call with the recruiter to go over my background and the role. After that, there was a technical screen with some coding problems. Then I had a few rounds focused on my previous ML work and how I tackle problems. The final round had some values questions and conversations about safety in AI. Everyone I met was sharp and respectful.

First there was an automated coding screen where you write code to satisfy test criteria, with no AI tools allowed. After that there was a similar coding screen but with an employee present. Finally there was a remote panel interview that had three technical interviews, one role-related discussion interview, and one culture fit interview. The questions were difficult but fair, and kind of ironic for an AI company, since the ban on using AI tools means you probably want to practice coding problems without AI. Using google is allowed, but I had a hard time because I hadn't used google for coding for about a year. The problems themselves are quite straightforward and AI would easily solve them.

I was reached out by a recruiter, and then they set up a first round tech screen that was a call with the hiring manager. Nothing special. Pretty standard flow. They told me to focus on concurrency for the coding part.