Stats/ML PhD Application Strategy Review - Fall 2025

[RockML]

Undergrad: Top 40 university, known for STEM

Major: Statistics

Minor: Computer Science

GPA: 3.8/4.0

Grad Level Courses taken: Statistics and Probability, Statistical Modeling, Machine Learning, Data Mining, Big Data Programming. (Sadly, no Real Analysis)

Research Exp: 6 months Machine Learning & Data Privacy research, get A score, no publication. 

 

Master: Top 10 university

Major: Machine Learning

GPA: 3.8/4.0

Grad Level Course taken: CV, NLP, LLM, Statistical Modeling, Optimization, Probabilistic Graphical Models

Research Exp: 1yr Deep Learning, LLM, GenAI research with CS Dept. Professor, 6 months Business Statistical Analysis (Regression, DiD) with Business Dept. Professor.

Publications: 3 conference papers were submitted to top ML conference, and at least 1 of them can be acpected before December.

 

GRE: not going to take it, I will focus on my current research.

Type of student: International male

Research Interests: Data Science, Machine Learning, Large Foundation Model, Generative AI, Data Privacy, AI for Healthcare (incoming).

Letters of Recommendation: From the three research teams that I worked with (as mentioned above), the undergrad one should be positive, the one from the CS Professor should be very positive, the one from the Business School Professor should be extremely positive.

Relevant Professional Exp: Staff MLE for a Start-up company, I led a team of 3 people to build an audio AI system. 1 intern as CV Enginner, 1 intern as NLP Engineer.

 

My thought: 

As I mentioned in the title, I’m seeking 'strategy' advice rather than a specific school list due to my unique background. My experience is at the intersection of Statistics and Computer Science. While I haven't taken Real Analysis, I have substantial experience in Advanced Algorithms, Advanced Deep Learning, Variational Inference, and ODE/PDE for Diffusion Models. Personally, I believe that mastering complex algorithms like PPO demonstrates my mathematical skills just as effectively, but I am not sure everyone can appreciate that.

So, my concern is am I still a good candidate for STAT PhD program (for top univ)? There're many remarkable ML researchers in STAT/OR department that I want to work with, and I pretty much enjoy STAT research vibe more than CS, I don't know, probably because CS student runs project all by themself and has less time to communicate with the mentor.

In addition, I am also thinking about working for 1 year and apply 2026 fall PhD programs. By doing so, I will at least have all my papers published, and if I have time, I can still research remotely with the CS Lab. It should be a huge improvement, if things go smoothly, I can have 3 top papers (include one paper as 1st author) rather than just 1 for this year.

Overall, I am looking for any advice to refine my strategy. If you have suggestions, even if not directly related to my question, I'd love to hear them as well. Thanks in advance, I'll try to reply all of the message and keep my application status updated.

 

 

Edited August 6 by RockML
typo