Statisics PhD Evaluation

[slt-]

Hey there guys, I'm currently completing my masters and I'm strongly considering applying to Statistics PhD programs next year after I'm done. My research interests are in statistical machine learning and optimization, although I've become increasingly interested in biostatistics as well.  

Undergrad Institution: Berkeley
Majors: B.A. Mathematics & Statistics (with Honors)
GPA: 3.81/4.00

Type of Student: Domestic Male

 

Masters Institution: Cambridge

MASt in Mathematical Statistics 

I'll have my grades for my masters sometime next year towards the end of the course. 

 

Type of Student: Domestic Male

GRE General Test:
Q: 170 
V: 169
W: 5
GRE Subject Test in Mathematics:
M: 82%

 

Programs Applying: Statistics

 

Research Experience:  

3 Research positions at Berkeley. All under EECS professors, one doing research in non-convex optimization, another in information theory and source codes, and one in statistical machine learning. I have two publications but none as a first author. I'll also soon begin conducting ML research under another notable CS professor here at Cambridge.
Awards/Honors/Recognitions: 

N/A
Pertinent Activities or Jobs: 

1 software engineering internship at a startup

Letters of Recommendation:

I'll probably get them from professors I've done research with so they should be really good. 

Coding Skills: 

Python, Java, R

Relevant Classes, Grades:  

Undergraduate Studies-

Mathematics- Calculus I,II,II(A, B, A), Linear Algebra and Differential Equations (A-), Discrete Math (A), Advanced/Abstract Linear Algebra (A), Abstract Algebra (A), Real Analysis (A), Complex Analysis (A), Measure Theory and Topology (A), Functional Analysis (A), Advanced Numerical Methods for Matrix Computations (A)

Statistics- Probability Theory (A), Statistics (A), Computing with Big Data (A), Stochastic Processes (A), Statistical Learning Theory (A), Time Series Analysis (A)

EECS-Introduction to Programming (B+), Foundations of Data Science (A-), Optimization Models in Engineering (A), Machine Learning (A), Convex Optimization (A), Non-Convex Optimization and Approximation (A), Information Theory (A) 

Graduate Studies-

Measure-Theoretic Probability, Advanced Financial Models, Stochastic Calculus and Applications, Modern Statistical Methods, Casual Inference , Topics in Statistical Theory, Computational Game Theory

I'll have my grades for my masters sometime next year towards the end of the course. Although, I anticipate I'll get good marks in them.

 

 

Right now I'm not sure if I want to go into academia or industry, so I guess the "name brand" of the school matters for me in industry. Programs I'm thinking of applying to right now are Stanford, Berkeley, Harvard, UW, CMU and UChicago. Here are some of the concerns that I have-

1. I obviously also need to add some safety and match schools as well so I was hoping I could get some feedback on that.

2. I have a few concerns regarding some of my poor grades in more basic classes like intro to programming, Calc II, Intro to Linear Algebra and Diff Eq's. I took these classes in the first three semesters of college when my mom was dying from terminal pancreatic cancer and as a result I never really got a chance to study, I spent all my time with her while I could. Should I explain these circumstances on my application?

3. Another concern I have is maybe my lack of dedicated programming classes. Although I learnt data structures and algorithms on my own time (and in practice through my software engineering internship), I never took a formal course in the topic or any other systems CS class. Could/Will this hurt me?

 

 

Thank you for reading this super lengthy post. Any and all help is appreciated

 

Edited November 26, 2019 by slt-

[bayessays]

Assuming your grades from your master's are good, you can definitely apply to any school.  You'll probably get into top 10 programs but I would add some safer programs in the top 20 too.  

 

Don't worry about a B. Nobody will think anything of it with the rest of your record.  You clearly have some coding skills, so no, not have programming classes will not matter. 

[slt-]

18 minutes ago, bayessays said:

Assuming your grades from your master's are good, you can definitely apply to any school.  You'll probably get into top 10 programs but I would add some safer programs in the top 20 too.  

 

Don't worry about a B. Nobody will think anything of it with the rest of your record.  You clearly have some coding skills, so no, not have programming classes will not matter. 

Phew, that's a relief, thanks so much! Also,  I have a geographical preference for the Bay Area (although obviously I'm open to any location). I know that Stanford and Berkeley are by far the toughest programs to get into, but is there anything more I can do to optimize my chances for those schools?

[bayessays]

I'm not sure of the exact admissions data, but Berkeley has a biostat PhD with some very excellent crossover faculty like van der Laan that fit your research interests.  If location is that important to you, it might make sense to apply to the biostat program (although I think the statistics program is bigger, and you definitely have a shot at it, so I'm not sure whether this is a good idea. 

UC Davis has a good program and is sort of nearby, and UCSC also has some really good people, but if you're open to going further from the Bay, you don't need to go to programs ranked this low.

If you're staying in the US and want to go into industry though, you'll have good prospects coming from any decent PhD program, at least in the technology industry.

[Fred210]

On 11/26/2019 at 11:44 AM, bayessays said:

I'm not sure of the exact admissions data, but Berkeley has a biostat PhD with some very excellent crossover faculty like van der Laan that fit your research interests.  If location is that important to you, it might make sense to apply to the biostat program (although I think the statistics program is bigger, and you definitely have a shot at it, so I'm not sure whether this is a good idea. 

UC Davis has a good program and is sort of nearby, and UCSC also has some really good people, but if you're open to going further from the Bay, you don't need to go to programs ranked this low.

If you're staying in the US and want to go into industry though, you'll have good prospects coming from any decent PhD program, at least in the technology industry.

I would also highly recommend applying to Berkeley Biostatistics with van der Laan. He deals with state of the art statistical machine-learning methods (TMLE, SuperLearner, HAL) that are rigorously backed by mathematics. Your very strong mathematical background will definitely stick out in the application, and I think you have a very good chance of getting in.