Stellar Code, Simulations Help Ye Land Big Tech Internship
Heidi Opdyke
- Associate Dean of Marketing and Communications, MCS
- Email opdyke@andrew.cmu.edu
- Phone 412-268-9982
During a discussion with an Amazon recruiter, Carnegie Mellon’s Runqiu Ye didn’t expect to be talking about binary stars. But the recruiter wanted to hear all about them.
Binary stars — two stars orbiting a common center of mass — impact almost all aspects of astronomy, from shaping galaxies to making black holes. And it turns out, working with code that simulates how the pairs of stars evolve from birth to death is directly relevant to interning at Amazon.
“The research experience I had definitely goes beyond physics with respect to essential skills like working with others, collaborating on a large code base and making contributions to the open-source community,” said Ye, a rising junior in computer science.
Ye spent the summer of 2024 writing code for an astrophysics simulation that tracks the mass, radius, temperature and orbits of millions of binary stars, including any interactions that happen between each partner star. Assistant Physics Professor Katie Breivik said that Ye’s ability to learn a new programming language and produce results quickly likely helped him stand out for the internship.
“The bulk of our simulation’s code is written in Fortran, which is often viewed as old or outdated and certainly not used much outside of scientific computing. Runqiu’s ability to get results even though he hadn't been directly trained is an awesome indicator of his willingness to figure things out,” said Breivik. “He stands out as one of the most driven students I've worked with at his stage.”
Ye said physics was a longtime interest for him, and he wanted to conduct physics research as an undergraduate. He approached Breivik after attending a research talk she gave, and Breivik encouraged him to apply for a Summer Undergraduate Research Fellowship (SURF) grant from Carnegie Mellon's Office of Undergraduate Research and Scholar Development.
Ye’s project started by comparing Breivik’s computer models of binary star interactions with actual observations of a recently discovered set of white dwarf binaries. The observations of the stars’ orbits as they revolved around each other didn’t quite match the models. To determine why, Ye ran simulations of millions of binaries to map the properties that influence the two stars’ orbits.
After three weeks, Ye developed a fine-grained grid of simulations.
“At this point it was clear that Runqiu really knew what he was doing with code and computation,” Breivik said.
Ye then started working with a new version of the code. His job was to add new calculations to the codebase that include detailed stellar physics properties to see if these upgraded models better matched observational data. He worked with a subset of code that involved the common envelope, a phase in a binary system’s life when both stars become wrapped in the same material. The common envelope phase is a not-well-understood step in the complex chain of events in binary stars’ lifetimes.
By generating artificial populations of stars in their model, Breivik said they can incorporate different assumptions for how interactions between the stars — like when and how the common envelope develops — can alter both the stars and their orbit.
“There are these huge data files about properties of the star at different stages. One of my main contributions was to add code to interpolate the common envelope parameter,” Ye said. “It’s a pretty large code and I needed to understand what all the different parts of code were doing so I could make my changes. That was quite an interesting experience.”
Research skills like these translate well to roles in the corporate sector.
“Analyzing an issue, forming a hypothesis and prototyping a solution — these critical analytical and problem-solving skills are in high demand, especially at leading companies like Amazon,” said Kevin Monahan, associate dean and director of Carnegie Mellon’s Career & Professional Development Center.
It was Ye’s first time contributing to a large code base.
“It was a great experience to have not just in physics, but also in developing and making contributions to the open-source and broader research communities.”
Ye said he is looking forward to translating his new skills to a different code base — this time at Amazon.
