AI Research Scientist · Onsite · Multiple Types

Meta — AI Research Scientist ✅ Passed

Level: Intern

Round: Onsite · Type: Multiple Types · Difficulty: 6/10 · Duration: 90 min · Interviewer: Unfriendly

Topics: Research, AI Coding, Algorithms, Fisher-Yates Shuffle, React Loop

Location: Paris, FR

Interview date: 2026-01-09

Summary

Interview Rounds Overview

• Round 1: Research
• Round 2: AI Coding

Details

My interview for the Meta MSL Paris team RS intern position consisted of two rounds, each 45 minutes long.

Round 1: Research

This round started with 5 minutes of small talk. I presented one of my research papers (the interviewer chose from two options). The key is to select a project/paper closely aligned with the team's focus. The interviewer asked questions about the project. The final 20 minutes were an open question: how would I build a team to develop an agent, given sufficient resources? I used the last 5 minutes to ask about the team.

Round 2: AI Coding

This round focused on AI coding, not using AI to write code. The interviewer (from the sister team of the first intern's host team) stated that questions could involve LeetCode problems, algorithms, or basic AI coding. No self-introduction was needed due to time constraints (45 minutes is expected for coding).

I was asked two questions:

1. A problem with many confusing conditions that tested O(N) implementation of random without replacement (Fisher-Yates shuffle). I was able to answer it correctly because of my computer science background, previous experience as an algorithm course TA, and having created a similar question before.
2. Implement a React loop based on provided APIs and functions.

I received the offer three days after I used a GDM offer to expedite the process. One week had passed without any response before that.