Skip to main content


Office Location

805 Columbus Avenue
577 Interdisciplinary Science and Engineering Complex (ISEC)
Boston, MA 02120

Mailing Address

Northeastern University
ATTN: Brett Daley, 435 ISEC
360 Huntington Avenue
Boston, MA 02115-5000

Research Interests

  • Machine learning


  • MMSc in Global Affairs, Tsinghua University, China
  • BS in Electrical and Computer Engineering, Northeastern University

About Brett

Where did you grow up?

Mendham, New Jersey.

What are the specifics of your educational background?

I completed my undergraduate degree at Northeastern. The co-op program undoubtedly had a major influence on my decision. While important career skills like professionalism, resume writing, and interviewing must often be learned on the job, they are built into the curriculum and embodied in the culture here. I found that very appealing. I was also excited by the prospect of living in a vibrant city like Boston, which was a refreshing change of pace from my small New Jersey hometown.

This past year, I studied abroad at Tsinghua University in Beijing, China. I was especially drawn to the Schwarzman Scholars Program, which provided me with an opportunity to meet people and explore disciplines outside of my field. In addition, China’s unique cultural differences, rapid economic development, and enormous manufacturing sector make it a fascinating case study for machine learning applications.

What are your research interests?

I specialize in a subfield of machine learning called deep reinforcement learning, which combines approximate models of the brain (“neural networks”) with an incentive mechanism analogous to a dopaminergic reward system. This allows the computer to learn how to perform a task through trial and error, rather than needing to be given explicit instructions to execute.

My particular research seeks to make this type of learning more efficient in the presence of rich sensory information, such as high-resolution images. Roughly speaking, providing richer data to the computer will cause it to learn more slowly — or not at all — as it is “overwhelmed” by a stream of extraneous details irrelevant to the task at hand. In contrast, the human mind can extract with ease the pertinent essence of a problem in order to solve it. This disparity is extremely interesting to me. For my experiments, I typically test new algorithms by measuring how quickly they can learn to play arcade games or control simulated robots without any prior knowledge.

I discovered this area of research rather serendipitously. I began my career in more traditional engineering fields, like computer architecture, circuit design, and firmware development. After working at Tesla and witnessing its swift progress on Autopilot and its highly automated Gigafactory in Nevada, I acquired a strong interest in machine learning, which led to my current research path.

What are the specifics of your industry experience?

I completed four co-ops while I was an undergraduate at Northeastern. I began at AMD designing microarchitectural features of next-generation x86 processors. The next year, I relocated to California for three consecutive semesters: I prototyped printed circuit boards in a new R&D laboratory at Flextronics, developed battery firmware for Tesla’s large-scale Powerpacks, and wrote software at SpaceX to automate various functional tests for the avionics in Crew Dragon.