First interview:
What did you do in your B.Sc, M.Sc. and Ph.D.?
Have you ever used TTS systems of Nuance? Or any other TTS or speech recognition system?
What part of the task do you think the most difficult one?
What are the steps that you see to do the job? (The jobs was working on TTS system on a specific language, my mother tongue, i.e. I was supposed to talk about text normalization, sound to phoneme, etc.)
Second interview:
Tell me about yourself (I talked about my scientific background, but at the end he said that's fine but I meant about yourself, your personality, etc.).
Could you explain about the phonetics of your language? The number of vowels, consonants, front or end vowels, etc.
How will you model the sound to phoneme part of the job? rule-based or statistically motivated or both?
What's the stress pattern like in your language? Is it easy to model? Is it predictable? Is it similar to English? etc.
What is allophone?