Research in Rochester: Rajaoberison helps explore the universe with machine learning
One of Heriniaina Rajaoberison’s high school tenets was to do hard things and take risks. He practiced this philosophy throughout his time at UR, which led to many exciting opportunities, one of which was working on NASA’s James Webb Space Telescope. Rajaoberison, a senior graduate in optical engineering, is chief technology officer and co-founder of Advanced Growing Resources, a UR student start-up developing optical sensing technology.
When Rajaoberison came to UR in his freshman year, he was interested in studying engineering – he just didn’t know what kind. The day before registering for classes, Rajaoberison took part in the departmental student fair. “There were some cool displays from the optics department, one was something playing tricks with your vision and stuff, and that was mind blowing to me,” Rajaoberison described. A professor told Rajaoberison to attend the first lecture of an optics class, and after attending, he wanted to know more. Rajaoberison took a few more courses in optics and then never left.
Rajaoberison became involved in research during his freshman year after meeting his optical advisor, Professor Jennifer Kruschwitz. Rajaoberison wasn’t sure what interested him, so Kruschwitz suggested a few labs he could contact. After a few meetings, he landed in Professor Jannick Rolland’s lab, where he worked on mixed reality optical design using Microsoft HoloLens.
The summer after Rajaoberison’s freshman year, he interned at the Laser Energy Laboratory, where he got a job thanks to Kruschwitz. “[Professor Kruschwitz’s] her husband, Professor Brian Kruschwitz, who works at LLE, was looking for someone for a summer project,” Rajaoberison explained. Apparently the project was supposed to be for someone in first or senior year, but I got lucky thanks to my connection[s] to be able to talk to him. I was really excited about the project, and it was my first time doing machine learning research. »
Rajaoberison found his passion for machine learning that summer and, through his LLE background, was able to use his machine learning skills with Professor Jim Fienup, who worked on the James Webb Space Telescope. Rajaoberison is an undergraduate research assistant in Fienup’s research group, where he works with Fienup and Ph.D. candidate Joseph Tang on wavefront sensing, using machine learning, for fine alignment primary mirrors of the James Webb Space Telescope.
“I learned a lot through the experience of Professor Fienup, who I would probably say is the godfather of phase recovery,” laughed Rajaoberison. “He helped me when I got stuck and supported me when I didn’t know how to do something. Joseph Tang also gives me a lot of advice on research and machine learning.
Rajaoberison writes machine learning algorithms to help James Webb’s 18 primary mirrors position themselves and counter piston phase error. Rajaoberison used synthetically gathered data from a computer simulation that attempted to determine how light would be seen through the telescope. This data would help determine how to position the telescope’s mirrors and overcome piston phase error, which impacts images taken by the James Webb Space Telescope.
“It’s like saying, NASA depends on what I’m doing here to get a really sharp picture,” Rajaoberison said. “Because what I’m doing is really helping the James Webb Space Telescope get a sharper image, and getting a sharp image means we can better understand the universe and where it came from.”
Last January, Rajaoberison, Fienup and Tang presented their work to NASA. Rajaoberison was the one who gave the presentation, an experience he found amazing as an undergraduate student. “It’s been amazing, I guess. I never thought I would have the chance to collaborate with NASA, and especially as a non-US citizen,” Rajaoberison said.
Rajaoberison encourages students interested in research to speak to their faculty advisors. “If you don’t know where to start […] speak with the academic advisor,” Rajaoberison said. “I was fortunate to have a wonderful counselor in Professor Jennifer Kruschwitz. She is awesome and [advisors] are willing to discuss with you what interests you.