Katarina: Navigating Computer Science, Research, AI's Future

When I first connected with Katarina Cheng, a fourth-year computer science and engineering undergraduate at MIT with a math minor, I wasn’t sure what to expect. Over our Zoom conversation, however, it became clear that she is deeply immersed in cryptography research, passionate about teaching, and reflective about the challenges of both research and academia.

Her journey to MIT, like many students there, started with a strong passion for math and science. But as she would later discover, her interests shifted from pure math to the more applied world of computer science.

"I was actually more interested in math early in high school," she told me. "MIT was my goal, but I initially thought I'd be doing more theoretical work. Over time, I got more involved in CS, and I realized I wanted to work on problems with more real-world applications."

That shift would eventually lead her to one of the most fascinating subfields in computer science: cryptography and security.

Curious about other women in STEM?

I spoke to several of them. It's worth listening to what they have to say.

Check it out

Katarina Cheng

The Transition to MIT: More Than Just Skipping Classes

The first year at MIT is known to be intense, even for the best students. I asked Katarina what that transition was like for her.

"I think I was overly ambitious my first semester," she admitted. "I skipped some intro classes and took more advanced ones. It was a good learning experience, but also a challenge. Coming from high school, I was used to getting everything right on tests. At MIT, I had to adjust to the idea that sometimes, I just wouldn’t be able to solve everything."

Skipping classes is a common decision among MIT students who want to get ahead, but Katarina regretted it—not necessarily for the content, but for the missed social connections.

"A lot of my classmates were in those intro classes, and I think I missed out on forming relationships early on," she said. "And in CS, some of the foundations would have been really useful before diving into harder material."

Despite the initial challenge, she quickly found her footing—especially when she took a cryptography course in her sophomore year that completely changed her trajectory.

"That was the moment I really fell in love with the field," she told me. "Since then, I’ve been trying to pursue cryptography and security in all my work."

Cryptography and Private Information Retrieval

Currently, Katarina’s research focuses on private information retrieval, a subfield of cryptography that explores ways for users to access online data without exposing their queries.

"Think about how we interact with the internet today—searching, retrieving data. All of that involves trade-offs between privacy and usability," she explained. "My work is about trying to redesign how we query databases in a way that prioritizes user privacy."

This research is part of a broader movement in computer science to rethink internet security and user protection. She hopes that her work can contribute to more secure systems without sacrificing efficiency.

But working on open-ended research at MIT was a very different experience than the structured projects she had done before.

"I was used to projects where you have a clear goal, like ‘implement this algorithm’ or ‘analyze this dataset,’" she said. "But in research, especially cryptography, it’s often open-ended. We started with a vague problem like ‘how can we make these protocols more efficient?’ And from there, we had to just explore. It was really terrifying at first."

She admitted that one of the biggest challenges in research is failure—trying different ideas, only to realize they don’t work.

"You try one thing, it doesn’t work. You try another thing, and that doesn’t work either. That was really hard to get used to," she said. "But eventually, you realize that’s just part of the process."

The Role of Collaboration: Researching in CSAIL

Katarina is currently part of CSAIL (Computer Science and Artificial Intelligence Laboratory), one of the most well-known research labs at MIT.

"CSAIL is huge—it covers everything from cryptography to AI," she explained. "My specific group is the Parallel and Distributed Operating Systems group, but my professor focuses more on security."

She got involved in research the way many MIT students do—by taking a class with a professor she admired, then reaching out to ask about research opportunities.

"I took his cryptography course, really liked it, and then just emailed him to ask if he had any open research spots," she said. "And it worked out."

Her research process involves a lot of discussion, reading, and thinking—often more than actual coding.

"People think CS research is all programming, but a lot of it is actually brainstorming, reading papers, and writing proofs," she explained. "For this project, we mostly worked at a whiteboard, throwing out ideas and seeing what sticks."

Teaching and Mentorship: "The TAs Make the Class"

Beyond research, Katarina has also spent several semesters as a teaching assistant (TA) for MIT’s algorithms and security courses.

"I really enjoy teaching," she told me. "I started tutoring in high school, and at MIT, I wanted to see if I liked TAing. It turned out to be one of my favorite experiences."

Teaching not only helped her understand the material on a deeper level but also gave her a valuable lesson in communication.

"My first semester TAing, I had no idea how to present material," she laughed. "I struggled with balancing intuition versus formality. And sometimes, I’d say something, and the students would just stare at me blankly."

Over time, she learned how to guide students without just giving them the answers.

"In office hours, sometimes students don’t know where they’re stuck. Learning how to help them think through a problem instead of just giving hints was a big learning experience for me."

The Future of AI, Cryptography, and Security

Given her interest in cryptography, I asked Katarina how she sees AI impacting security and privacy.

"One of the biggest research areas right now is securing AI models—how do we train them without violating user privacy?" she said. "AI is trained on massive datasets, often without people’s consent. Cryptography can help ensure that this training is done in a way that doesn’t expose private data."

As for whether AI poses a threat to cryptographic security?

"Not really," she said. "Cryptography is mathematically provable. Unless AI somehow figures out how to break the most fundamental assumptions—like factoring massive numbers efficiently—we’re safe for now."

However, she does see a major shift in AI regulation.

"Unlike the early internet, AI is being regulated very quickly," she pointed out. "The EU already has an AI act in place, and we’ll have to see how that affects the industry moving forward."

Looking Ahead: PhD and Beyond

With graduation approaching, Katarina is currently applying to PhD programs in cryptography and security.

"For the next five or six years, I’ll probably stay in academia and keep doing research," she said. "Beyond that, I don’t know yet—maybe something applied, maybe industry. But I want to build a strong research foundation first."

If she could give advice to younger students, she’d tell them to try more things and take risks.

"In high school, I stuck with what I was good at. I wish I had pushed myself outside my comfort zone earlier," she reflected. "Even in CS, I was afraid to take coding-heavy classes because I wasn’t the best programmer. But once I forced myself to do it, I grew a lot."

Her biggest takeaway from MIT?

"Explore as much as you can. Find what excites you. And don’t be afraid to fail—it’s just part of the process."