Synthesia Machine Learning Engineer Interview Questions
& Process

The hiring process started with an HR outreach, followed by a first interview with the team lead. Then, I was given a take-home task, estimated to take 6 hours, involving a decoder-based language modeling task with an additional dimension. Unfortunately, the HR mismanaged the timeline, providing me with less than 3.5 hours. I spent half an hour understanding the task and code, and over 40 minutes setting up the local environment, leaving just over 2 hours for completion. The feedback indicated my solution was reasonable, but they had only one position, and another candidate was already favored. It was a tough task but doable with adequate time.

I had a phone call with the CTO and Head of Talent, it was a really good first interview. Then I got a timed home assignment which was pretty interesting and showed off my skills well. After that, I had a debrief with two colleagues, one senior and one junior. They asked a bunch of relevant questions about the implementation choices I made. Even though I answered everything well and got great feedback on the call, they decided not to move forward and didn't give much of an explanation.

After getting a referral, I spoke with a Synthesia recruiter who told me about the company and encouraged me to apply. Then, I had phone interviews with the head of research and some potential teammates. Following that, I completed a take-home technical test, which was a substantial amount of work. The process concluded with a call to discuss the take-home test. It was a good experience overall, with consistent communication.