ePacket Interview 08.19.2014

• ePacket which I thought is a normal start up as usual, however it is not at least from my perspective. Actually it's more like a very potential technology company founded by top university students. Of course it is still a start up but a very unusual one. 
• The interviewer is a Stanford student in EE and obviously he found no interests on me.
• On the one hand I downplayed this interview just treated it as a trial for my full time interview. In fact I'm not ready to start and this interview leaded from a mistake application which I original wanted to apply internship. I didn't know what this company do even before one hour ahead of the interview
• On the other hand, maybe even if I'm fully prepared both in skills, projects and interview skills, he would still have no interest on me because he is scouting someone much more smart and at least at a top university like Stanford.
• But no matter the result is, it doesn't matter, what matters is what I learned from it

1. As usual, start with a self description

1. better put algorithm and data structure ahead of the low level knowledge which actually I cannot say I know a lot. Don't say anything like what you are interested in or what you are good it if you cannot convince interviewer by evidence and facts.
2. Must review of last quarter's advanced architecture class. It deserves it.

1. What field you mainly focused on during your college

1. The algorithm and data structure is the only field I'm not afraid of being ask any question, so talk it as much as possible and dominant the conversation
2. This question needs still consideration
3. If you said you are interested in architecture or network, it's very likely that be asked about what have you done or researched on. So if just one project, I'd better don't say like that.

1. Talk about your projects experience. What they want to hear is what difficult problems you solved or what improvement you did rather then plainly describe what you did or what you implemented. This is very important. Don't make your description boring without any useful information

1. First at all, review what has been implemented and what has been solved in each project
2. Package each project, extract the featured and highlighted part. Think about how to show them
3. For the system kernel, the obvious key is to show how did I debug the program. If said that the original simulator is not powerful enough, it is very likely asked that did you improve that?
4. For the PIMDM show the architecture of the job, the complexity is the featured point of this project
5. About the database management system, try to think out what problems you have solved. 
1. For example today I was asked that in the B+ tree layer if the data store both in memory and disk, the time latency is significant, how do you solve it. 
2. Review about the bug that runs different result on mac and linux about the memcpy error leading from the address overlap
3. What they want to know is not plainly description. They want to know the stuff exactly like this.

1. Asked about my research experience

1. Shit answer

1. Asked about explaining the most proud project or maximum project is size and what is your role that what you did in the project

1. my answer is the database system. However I just cannot clearly specify what is my role because the fact is that we almost did the project together and separated the project evenly. This needs consideration.

1. At last asked if I have any question about the company