ECE faculty member’s lab benefits all involved
MU electrical and computer engineering Assistant Professor Michela Becchi recently received the UM System’s President’s Award for Early Career Excellence in no small part because she is the top National Science Foundation-funded faculty member on the MU campus.
Becchi told her graduate research assistants that her success is a reflection of the work they do in her lab, and they say their success has everything to do with their excellent mentor. Two of Becchi’s graduate assistants are nearing the end of their programs. Da “Daniel” Li, is set to defend his dissertation this summer, and Michael Butler will defend his thesis this summer. The pair has landed dream jobs at Facebook and Google, respectively.
During his last two years of college in Beijing, China, Da “Daniel” Li focused his efforts on parallel computing, an area of study that gained prominence with the emergence of many-core processors in the last dozen years. Taking graphics processing units (GPU) as an example, these electronic circuits, originally designed to accelerate the performance of computer graphical applications, have since been tapped to expedite complex data analysis by “stringing” their computing power together.
“Accelerating applications on many-core devices is more and more important because of [data-intense research applications in such things as] bioinformatics, social network analysis and machine learning,” Li said. “There is lots of exciting work for things like self-driving cars being done based on deep learning using GPUs.”
In 2010, Li chanced to read a paper written by Becchi whose research is based in parallel computing. He downloaded her open source software and used it for his research. The following year when Li planed to apply graduate school, he noticed on Becchi’s homepage that she was relocating to the University of Missouri. He reached out to inquire about joining her research team and was accepted.
Butler arrived at MU in the fall of 2010 as a freshman. He joined ECE Professor Marge Skubic’s research group his junior year and when it came time to choose a mentor for his master’s degree, he said he decided to work with Becchi because of her parallel processing course.
“There was something about the way she taught it,” Butler said. “She went into high-end systems in an intuitive way that was not threatening.”
He joined Becchi’s lab and is very pleased with his decision.
“Dr. Becchi knows exactly what the lay of the land is in the current field,” Butler said. She makes time for you; she helps you; she guides you; she mentors you. She gives advice that actually helps you and she’s easy to talk to.”
Ph.D. candidate Li said he has been working to accelerate emerging applications on many-core processors, such as Nvidia GPUs and Intel Xeon Phi.
“I am looking at problems in an application domain and then working to build a bridge between the application and the parallel hardware to make it work more effectively. Dr. Becchi is very helpful. She has good suggestions about the research,” Li said and added, “She also encourages everyone in our lab to do internships.”
The doctoral student has done internships at NEC Labs in New Jersey, at AT&T Labs and last spring, he worked at the Sony U.S. Research Center. In addition, he applied for and was granted a $6000 scholarship for Ph.D. students abroad from the Chinese government.
When Li started looking for a job, he was interviewed by and received offers from Amazon, Apple, Facebook and Google. He said he chose Facebook because, for him, it offers a better career because as a research scientist, the work will be a mix of engineering and research.
“It is a relatively young company and still growing,” Li said. “The code [he will be working with] will impact billions of users immediately. And at Facebook, one engineer serves more than one-million users everyday.”
Butler said his area of research has been to try to make multicore processing faster and more efficient for the programmer especially in the case of sharing data between CPUs and GPUs. He completed internships at Garmin and NEC Labs and will be working on the Android development team at Google starting in August.
“I’ll be working on a programming framework called Renderscript that works on computer vision systems,” said Butler. “It’s an API [application programming interface] that simplifies parallel programming and can be used for both CPUs and GPUs.
“Everything they are trying to do is what I have been working on in my research,” said Butler.