This interview was for a computer vision engineer. I worked with the HR department to schedule a phone interview.

The phone interviewer asked me about the LoG function, corner detection, and their time complexities. He seemed very capable and accomplished, given his knowledge.

I was invited for an onsite interview and had five rounds, including a lunch interview.

Round 1: A challenging interview with a principal engineer who knew his field well. He asked questions about OpenGL, OpenCV, homography, and SIFT in detail.

Round 2: I spoke briefly about myself, then was asked to develop an algorithm and code to find the index of a number in a 2D array sorted in ascending order both horizontally and vertically.

Round 3: Lunch interview. Many questions about object-oriented design. I was asked to print numbers in hexadecimal and about the overhead of a class (including methods).

Round 4: The hiring team lead asked a series of behavioral questions. He then posed a couple of easy questions: one to twiddle nibbles in a number, and another to insert a string within another.

Round 5: The final interview was on computer vision, and I was asked about RANSAC, k-d trees, SIFT, homography, and so on.

During rounds 2, 4, and 5, I explained my PhD topic, but the other interviewers were not greatly interested.

How do you calibrate a camera to find extrinsic parameters?