Edulix
PhD in systems - Need some advice - Printable Version

+- Edulix (http://www.edulix.com/forum)
+-- Forum: II. Profile Evaluation Phase (/forumdisplay.php?fid=37)
+--- Forum: PhD Forum (/forumdisplay.php?fid=66)
+--- Thread: PhD in systems - Need some advice (/showthread.php?tid=183536)



PhD in systems - Need some advice - sajinsabu - 07-15-2017 02:40 AM

I do not know whether this is the right place to ask. If not, please redirect me to the right forum.

I am a computer engineering student at UNC Charlotte, specializing in operating systems. Recently, I realized that I have a deep urge to explore more in systems, and academia or R&D job will suit me more than any regular job.

My primary area of interest is systems (Distributed systems and OS).

My current profile is as follows -

Academics -

Bachelor's - B.Tech Electrical engineering (Univ of Kerala)- CGPA- 7.34/10 - bang average stuff done
Master's - MS Electrical (UNC Charlotte) - GPA 4.0 - did decent work (will give below) - will have good recommendations.

Work

Intern at Wind River Systems - Working on their VxWorks kernel - really high quality kernel work - (2 months)

IBM India (1.5 yrs) - nothing productive done on a boring IT job.

Papers - (Nothing related)

1. A Novel Brain-Controlled Multi-Fingered Dexterous Prosthetic Hand (Under Review) - Journal of Robotics
2. Innovative Low Cost Mars Flyby Spacecraft for Safe Interplanetary Human Mission – The Mars Society

Projects -
1. jOS exokernel (MIT 6.828): The project was split into 6 major parts that build on each other, culminating in a primitive operating system on which you can run simple commands through your own shell and will have Unix-like functions (e.g., fork, exec). Culminates with the development of an Ethernet device driver from scratch.

2. NinerSense Home Security system: Designed and prototyped an end to end IoT system using Watson IoT Foundation (Bluemix) to collect and analyze data from various sensors deployed throughout an apartment. An Android app was created as a remote endpoint to monitor the sensors and get live feed from the webcam.

3. Implementation and Performance Evaluation of Convolutional Neural Networks – Implemented a convolutional neural network on CPU and GPU. On the CPU, AVX intrinsics optimizations, cache blocking, and OpenMP optimizations were used. On the GPU, shared memory implementation was done using CUDA.

4. Cinema Booking System using Docker, and Nodejs Microservices – Currently working on developing a model cinema booking microservice using node.js, and deploying it to docker swarm cluster.

5. Priority Based Pre-Threaded Image processing Server - Developed an image processing server that interacts with clients through a pool of worker threads using the producer consumer model.


I have looked at the systems research at a lot of universities, and found that my choices are plenty. I even received a positive reply from Dr. Daniel Wong from UC Riverside.

I have very little idea on what my chances are in the really top schools with good systems research. For example, UC San Diego, Virginia Tech, University of Virginia, Wisconsin, Univ of Washington, GaTech etc has top quality work going on.

Can anyone help me in knowing how high I can aim, because I am completely clueless on that. I want to apply to a maximum of 3 schools, and want to choose professors based on knowing how high I can aim. I will make a separate profile evaluation post then.

PS- I don't know whether I will be taking a thesis, and time is running out. I wanted to do one, but I didn't find any work going on that aligned with my interests at UNCC.


Tagging
@[Edulix_Editor]
@[Rogue]
@[The_Observer]
@[The_Chharraa]


RE: PhD in systems - Need some advice - Edulix_Editor - 07-15-2017 03:28 AM

High @[sajinsabu]

Give me a day and I will give you some key pointers. We love it when our Edulixians come back to Edulix for PhD, job, etc questions.......

More later...........

EE


RE: PhD in systems - Need some advice - sajinsabu - 07-15-2017 06:24 AM

Thank you @[Edulix_Editor]

I will be here on Edulix probably all summer trying to help some applicants out! Very Happy


RE: PhD in systems - Need some advice - The_Chharraa - 07-15-2017 08:07 AM

Hi,

Great to see you're prepping up for PhD. Smile
I'm not good with the fields you work with. Nonetheless, I've been in research environment for a decent duration that could allow me to give few suggestions. Take them with a grain of salt:

1) I too have PhD in my mind so as and when I get chance I try to asked or ask my PI if they check or reply to e-mails they receive. Reply is mostly no. They throw them to trash.

Second point negates first point but read it.

2) There are conflicting views for prof-letting in this sub. But based on my experience and conversation from colleagues I think prof-letting does help but only to a certain extent. With me it has worked if I've asked person if he/she would be hiring for Fall-XX session and if I'm interested to work with them then which department should I submit my application. If they are looking for folks they would reply immediately. If not then, it will go unheard.
Again, prof might be traveling and didn't have time to reply and might see your application later on. So, worth going for this (prof-letting) approach.
Or, at times prof would say my lab works in applied physics and we are looking for such and such. This gives you a good idea, some times.


RE: PhD in systems - Need some advice - Rogue - 07-16-2017 03:23 AM

Unfortunately I happen to have very little experience with PhD applications in the US. I believe the other senior members you have tagged will be very helpful in this regard.

Good luck mate! Hope you find a great PhD program in the US. If you are considering something in the EU on the other hand feel free to buss me for suggestions and guidance.

Cheers!!


RE: PhD in systems - Need some advice - aditya57 - 07-16-2017 07:16 AM

As someone who took the same path as going from MS to PhD in the US, my top tip is to apply to schools post contacting the PI you want to work with. Since you are not divesting to many schools and not sure of the schools, prof-letting is really important in your case. For reasons of full disclosure, I haven't done that and that is my regret in the application cycle.

It is somewhat easier for professors to respond to students from US universities as the mail won't be discarded as spam right away. What you want to discuss with the professor is the black box, you need to come up with. Hope that helps.


RE: PhD in systems - Need some advice - sajinsabu - 07-16-2017 09:29 AM

(07-16-2017 07:16 AM)aditya57 Wrote:  As someone who took the same path as going from MS to PhD in the US, my top tip is to apply to schools post contacting the PI you want to work with. Since you are not divesting to many schools and not sure of the schools, prof-letting is really important in your case. For reasons of full disclosure, I haven't done that and that is my regret in the application cycle.

It is somewhat easier for professors to respond to students from US universities as the mail won't be discarded as spam right away. What you want to discuss with the professor is the black box, you need to come up with. Hope that helps.

Thank you @[aditya57] Smile
In fact, I got positive replies from 2 of the three professors whom I emailed, one from UC Riverside and other from UC Irvine.
Both of them have exciting research going on, so I'm stoked Smile
I want to see how high I can go. Hoping to get in to a big name university, under a great professor.


RE: PhD in systems - Need some advice - Gaussian - 07-16-2017 03:10 PM

@[sajinsabu] I wish I could help you here but I am not a systems guy. I would definitely follow what Edulix_Editor is going to suggest.

That said, let me tell you why I could not hold myself back from posting in your thread despite the fact that I have nothing as such to say. I am amazed and blown away by the list of projects you did. You would give some experienced software engineers a run for their money! I wish, rather I deeply wish, your publication side was equally strong. However, you have some other balancing factors like a masters degree from US, and hopefully some powerful recommendations. Try to play your cards well. Good luck!


RE: PhD in systems - Need some advice - sajinsabu - 07-16-2017 09:43 PM

Thanks a lot @[Gaussian] Smile


RE: PhD in systems - Need some advice - Edulix_Editor - 07-16-2017 11:40 PM

@[sajinsabu]

Ok, I had the opportunity to think through this. This is my feedback:

1. For PhD, it is almost a requirement to have MS thesis work. Personally, I will give more weight for someone who has already done a thesis work and published it or presented it in a conference. To my shock, I see data point after data point on the lack of correlation between high GPA and ability to do good research. Your projects are good indications; but not as good as a thesis which requires writing it and subjecting it to vetting at various levels.

I think you should aim for PhD in Fall 2018. That gives you almost a year to wrap up MS thesis and get some publication out. I strongly recommend that as it will really enable you to aim high.

2. With regard to list of programs where you could apply, I think you should aim higher and also identify the professors you are interested in.

The likes of UCSD, USC, VT, GT and potentially UW and others are not out as long as you have a good package which includes thesis some publication.

Equally, you should look at good professors in average programs as well. Some of the top cited professors are in average programs; but they produce highly rated research and mint great PhDs.

For example, Daniel Wong is a good example, He has potential. In his case, if I were you, I will go for Murali Annavaram at USC who is his advisor. He seems selective; but has good record in placing his PhD cards in academia & top notch R&D places.

3. With regard to prof-letting, my opinion is that - unlike MS - most of us pay attention to targeted, short email enquiries for PhD and post-doc positions. But don't send everything in 1 email. Just a summary with what you wish and some key links etc is good enough to get attention. Also, email/talk to some of the PhD students in the Lab. Some

In the case of those who are already in the US, I encourage to do some campus visits and stop by to visit the professor and his/her lab.

With this background, I think you can get started. I will be more than happy to walk you through the steps as you make decisions......

cheers,

EE


RE: PhD in systems - Need some advice - sajinsabu - 07-17-2017 10:17 PM

Thanks @[Edulix_Editor]

I will try to find an idea for thesis, and also publish a couple of papers Smile

Thank you so much!


RE: PhD in systems - Need some advice - The_Minuteman - 07-19-2017 07:19 PM

Here is my take on this (even though you did not ask me for it.. Razz)

If you want to apply to top schools, the two main things that matter are publication and letters of recommendation. Once you apply to top schools for a Ph.D, your work should be able to speak for itself. Remember, you are competing with an applicant pool where each person is bringing their absolute best to the table. I am talking from personal experience here. I had two guaranteed Ph.D admits from UMass Amherst and UNC Chapel Hill, but I wanted to just test the waters and see what would happen if I applied to some top schools.

When I applied, I had finished my thesis and had at least 3 papers in tier-2 conferences, one of them a first author paper. My GPA was not bad either. I got rejects from all my top tier school (which I knew), but why? I am sure it was because of the lack of publications (yes, 3 is on the lower side for top schools) and I am kind of certain one of my recommendations was not as strong as it should have been. That is water under the bridge now, but in hindsight, my profile was not what it should have been for a Ph.D at a top school.

If you want to get in to one of those schools, publish, publish and publish. Once you do that, your recommendations will sort themselves out due to your collaborations and from your advisor. The latter is just a function of the former.

Hope this helps.


RE: PhD in systems - Need some advice - sajinsabu - 07-19-2017 11:32 PM

That was helpful.
Thank you @[The_Minuteman]

I will have two more papers published in tier 2 conferences within 6 months, both as second author. But, they are not directly related to my field. Sad

But, anyway I have time on my side. So, will try to come up with something substantial soon enough!

Thanks a lot again Smile