I applied through a recruiter. The process took 2 weeks.

Interview process: I interviewed for a Meta summer internship, and the process consisted of two 45-minute live coding interviews. Each interview followed a similar structure: one easier warm-up problem followed by one medium-level problem. Both sessions were fast-paced. The interviewer expected me to clarify requirements quickly, propose an approach, and implement efficiently within the time limit. There wasn't much slack for long exploration, so it helped to communicate clearly, write clean code, and validate edge cases as I went.

Interview questions: The questions were standard data structures and algorithms style. The "easy" question typically checked basic implementation and correctness, while the "medium" question required a more structured approach and careful handling of edge cases and complexity. I won't share exact questions, but expect common DS/Algo patterns.

What went well: The interviews were structured and consistent (easy to medium). Interviewers cared about communication, explaining reasoning and trade-offs, and not just the final code. Complexity discussion and edge cases mattered.

What was challenging: Time pressure: two problems in 45 minutes means you need to move quickly. The medium question often required both correctness and optimisation, which meant thinking about time and space complexity early.

Tips for candidates:

• Practice solving two problems in 45 minutes to simulate the pacing.
• Focus on core DS/Algo patterns, such as arrays and strings, hash maps, two pointers, sliding window, stacks and queues, BFS and DFS, heaps, and basic DP.
• Get in the habit of clarifying constraints, outlining the approach, coding cleanly, testing a few edge cases, and stating complexity.

One of the key questions was a typical array/string problem solved with a two-pointer or sliding-window technique: maintain a window, update a frequency map/counters as the pointers move, and ensure the solution runs in linear time. I was also asked a graph-style BFS question that required using a queue, tracking visited states, and returning the minimum number of steps under constraints, with careful handling of corner cases.