Daniel Stein '18: 8 Steps to a Harvard PhD Program (AI Included)

Place of Influence: Harvard Medical School, Department of Biomedical Informatics

Role: PhD Student, Bioinformatics and Integrative Genomics

Focus: The study of immune responses in autoimmune diseases like psoriasis, with the goal of improving treatment outcomes and understanding why some patients relapse

 

8 Steps to a Harvard PhD Program (AI Included)

While at MA, Daniel Stein was known for his hard work, thoughtfulness, and humility. 

Today, as a PhD student in Bioinformatics and Integrative Genomics at Harvard Medical School, he puts those traits to work using advanced data analysis to understand immune responses in autoimmune diseases.

In particular, he researches illnesses like psoriasis and why some patients respond well to treatment while others relapse.

So what steps took Stein from MA to Harvard? Read on to find out!

 

1. Learn from Mom and Dad

Early on, Stein's parents nurtured his interest in science and math, setting up science experiments in the kitchen and teaching the foundational math skills that would serve him down the road.

2. See Challenge as a Friend

While at MA, Stein didn't confine himself to the International Space Station Research and math teams. He engaged in artistic and physical pursuits including cross-country, piano, and violin. After winning the 2017 SPPTA Concerto Competition, Stein explained: "practicing and performing beautiful pieces...brings wonderful challenge and delight." The recognition of challenge as a friend, not foe, helped set Stein up for success.

3. Explore Entry-Level Research

The summers after his sophomore and junior years, Stein joined research programs at UMN and MIT. The MIT summer program gave Stein his first encounter with artificial intelligence. "I used deep learning to interpret tumor mutations, which started my journey with machine learning and AI," he says.

4. Take the Online Course

Ever eager to grow, Stein took a free online course in artificial intelligence the summer before college. This bolstered his grasp on the technology and prepared him for step #5.

5. Take an Early Internship

Before his first year at MIT, Stein took a summer internship with VivaQuant, a provider of remote cardiac monitoring services. He wrote software that used AI to analyze data, remove noise, and help doctors detect and understand abnormalities. At the time, this was cutting-edge technology. Stein's supervisor, Marina Brockway, said: "There's nothing else like it on the market."* Stein was 18.

6. Tackle Real-World Problems at School

There is nothing quite like real-world application to cement your learning. Stein's undergraduate and master's work focused on computer science and engineering, as well as biological engineering. He continued to work with VivaQuant while in school. Among other things, he designed an FDA-cleared model for accurate and efficient identification of premature ventricular contractions (PVCs).

7. Ask: Where am I insatiably curious?

It's been said that one should be very careful in choosing a doctoral program: the average biomedical student takes 5.7 years to complete their research and dissertation, which is a long time to study one subject area. Stein decided to switch gears and focus on immune responses, which he finds fascinating because "you have all these different kinds of cells that take on special jobs. They talk to each other and work together to build this whole dynamic immune response." And so, in 2022, Stein began his current journey at Harvard.

8. Recognize your Role as a human

What will the future bring? Will AI take Stein's job? That depends, he says. "If I think of myself as a coder, well, at some point AI will streamline much of that process. I see my role as a researcher as much more than that. AI can free me up to focus on the important parts of the job, like thinking about what problems are valuable and worth pursuing, as well as evaluating the ethics and morality in research. I think that humans guiding the process is very important."

 

On Artificial Intelligence

Stein entered the world of machine learning well before ChatGPT and the wider public awareness of artificial intelligence. Here he shares a bit about his current experience with AI.

Favorite AI Uses 

I use AI to deal with mundane tasks so I can spend my research time in more interesting ways. I've used it to help me rewrite code, understand unfamiliar software, and as a search engine.

AI Impact on the Medical Field 

AI can identify patterns in biological data that humans wouldn't be able to spot themselves, leading to new drug development, advances in preventative care, and more efficient research. AI will likely lead to more efficient diagnoses based on medical imaging. 

Word of Caution

There's much work involved in showing that an AI model does what you expected it to do. Educationally, it's still important to learn how to do things and think for yourself­—without using AI as a crutch.