I entered Duke in 1992 with a plan to graduate with a triple major in
three years. Half way through that I burned out and had to take a
year off, I ended up as the sysadmin and tech support for an internet
access provider. I went back for a semester and then took a leave of
absence to start a company (with my ACM programming contest teammates)
that did real-time analysis of commodities data (similar to a hedge
fund except that it was for a single investor). After two years I
left that to finish my undergrad degree (simplified to two majors in 7
semesters, except spread over 6 years). My first college foray into
the ACM was the most successful, with a 3rd place finish at the
internationals. The next year hopes were very high and our team was
much stronger, but we flamed out. Later on I was the coach for Duke's
ACM team one year, and I focused mainly on making sure they were
relaxed and had a good time; I think they placed 5th.
After undergrad I rejoined my former teammates for more financial
number crunching in C++. The rest of the market eventually caught up
to our techniques (one way to look at the commodities markets is as a
way to get paid for releasing information to the world, so statistics
are only valuable if you are the only one that can compute them). Our
models worked for currencies too but the market itself was too
inefficient to make trading possible (any probabilistic profits were
overwhelmed by the transaction costs), so we decided to try to fill
the gap by building a currency trading platform that was more
efficient. That project turned into Forexster (www.forexster.com),
which is still riding the fine line between success and failure. It
is written mostly in Java, with a tiny bit of C++ in the cluster
interconnect.
In the course of Forexster I ended up reading a lot of research papers
about real-time database design and speculative concurrency control.
I realized that graduate school was about more than just designing new
data structures (I figured Knuth had that one covered pretty well) but
that it was about creating abstractions. Human brains only have a
finite size, so raising the level of abstraction translates directly
into increases in the size of the problem that can be tackled
successfully. Suitably motivated, I applied to grad school.
During college I started rock climbing, and it became part of my core
identity. I love Yosemite and the high Sierra, but I have a special
fondness for the North Carolina mountains.
I met my wife Paola at Duke during my senior year, we were married in
2003. No children yet. She decided to go for her PhD at the same
time as I go for mine (she already has a master's in public health in
biostatistics), so we both quit our jobs to become students again. We
live in San Francisco proper, she takes the BART to Berkeley and I
take the Caltrain to Palo Alto. We're both in our second year, but I
think that she will finish much more quickly than me.
Nate Bronson