Were you always planning to become an astronomer?

No. I went to graduate school in the physics department at the University of California, Berkeley, but I didn't know what I wanted to study. After my first year, I got a job with the Berkeley Automated Supernova Search, analysing images. I had no background in astronomy and didn't even know what supernovae were, but it sounded interesting because it was unexplored and would make use of my analytical thinking skills. Later, I began working on the distant-supernovae search, which ultimately became Berkeley's Supernova Cosmology Project. The distant-supernovae search did not achieve success until after I had left and started my postdoc at Fermilab in Batavia, Illinois, but later versions of the search benefited from the wrong turns we made starting out. Eventually the data were used in the discovery that the Universe is accelerating. Those findings led to the idea of 'dark energy'. My career has been shaped by a theme — although projects can seem bleak at the start, continuing to work on them can lead to an important result.

How were supernovae found?

First we had to work out what was going wrong. We couldn't have done it without constant funding for the supernova group from the US Department of Energy at Lawrence Berkeley National Laboratory — there has to be a way to keep going long enough to get things to work. You need to learn from your mistakes. One thing we learned is that to find a supernova, you can't record images of the sky once, and come back the next year and expect things to be the same. Too many things in the sky change to be able to tell which objects are supernovae and which aren't. We learned to capture an image of the sky before and after a full Moon to get the best spectra for finding supernovae.

Describe how an early achievement helped to chart your career course.

As a graduate student, I had the task of redesigning a filter wheel that was part of an instrument for the Anglo-Australian Telescope, one of the first Southern Hemisphere telescopes to offer high-resolution and computer-controlled spectrographs. I hadn't worked on hardware. I didn't know anything about optics or filter wheels, but I talked to the engineers and got answers. It came together, but the challenges showed me that with projects that push the envelope of what is known and possible, you are going to have to learn new things.

How did you forge the partnership with China?

It came together from both sides. I had been analysing data from the Sloan Digital Sky Survey in Sunspot, New Mexico, but data collection ended in 2008. So I was looking for a new project. I was interested in working on the Large Sky Area Multi-Object Fiber Spectroscopic Telescope (LAMOST) in Xinglong, China, because I wanted to continue with a galactic evolution project that I was exploring using Sloan data. But I needed data at a bigger scale, taking measurements on millions of stars. Using LAMOST I will have that, because it can take 4,000 spectra at once. I was also approached by the Chinese delegates as they talked to people involved in building Sloan, to learn how to make LAMOST successful. They contacted me because they wanted people to help them build the software.

What are the challenges in being a member of the first US team to join a Chinese-led astronomy project?

The Chinese structure for science is not similar to US or European structures. In China, individuals rather than teams are in charge. In US and European collaborations, committees are formed with representatives who have voting rights. It is like a democracy. With such different structures, it has been a challenge to define everyone's objectives, responsibilities and rights. The US National Science Foundation expects us to spell out the details in a proposal which might be funded up to a year later, whereas the Chinese organizations want to start working together and see how the relationship evolves, so there is a mismatch.

So have your goals changed?

No. No matter what the top layers look like, the scientists' expectations are similar. The challenge is getting the big organizations to recognize each other's systems rather than getting individuals to work together. I'll travel to China two to four times a year to make this programme a success.

What is your motto?

For the longest time, I told myself, “I can do anything”. When I turned 40, I realized that there are some things I simply can't do. Yet my motto got me pretty far before I figured that out.