Trump Card: Why a Pro Poker Player Bet on CUDA

by Brian Caulfield

You don’t quit your corporate job to become a full-time professional poker player just to get stuck in a classroom — with a massive student loan. That, as they say, would be a sucker bet.

Instead, Javier Torresola studies software engineering on his iPad from his home in Vancouver. Javier, 38, left his job as a thermal engineer at Intel in 2005 — and he’s been the master of his own time ever since.

Javier has just completed an online course through Udacity, ‘Introduction to Parallel Programming,’ to broaden his programming skills.

Javier Torresola left his corporate job to become a professional poker player.
Javier Torresola left a corporate job to become a pro poker player.

The course gives students a convenient way to learn about CUDA, the parallel programming model that harnesses GPUs to accelerate a broad range of scientific and commercial applications. “It’s really cool that I can learn parallel programming on my own time,” Javier says.

Those skills might even help him become a better poker player.

A Maverick Approach to Learning

Udacity, launched in January 2012, was co-founded by Sebastian Thrun – a Google Fellow who helped start the search company’s Google X Lab — to upend traditional ideas of what it means to learn in a connected era.

Udacity offers massive open online courses (MOOCs) — a term for courses that use the web to weave interactions between teachers, students and teaching assistants into self-paced classes that are free for anyone to take.

Udacity now offers more than 20 online courses for students to take at their own pace. A select group of tech companies, such as NVIDIA, Google, and Autodesk are working with Udacity to develop some of Udacity’s courses, aligning the interests of students, who want skills employers value, and the employers seeking to hire skilled workers.

Know When To Hold ‘Em

The “Introduction to Parallel Programming” (CS344) course, launched in February 2013, is taught by NVIDIA’s David Luebke and UC Davis’ John Owens. To make things interesting, NVIDIA sponsored a competition to rank how students performed on their homework assignments.

Javier was among the top scorers, getting a GeForce GTX 680 for his efforts.

The MIT-trained mechanical engineer’s primary goal: to acquire the skills he needs to become a freelance software engineer. His stretch goal: become an accomplished enough software engineer to begin simulating poker games so he can find strategies for playing with a technique called ‘fictitious play.’

“At the end you can get a full strategy for the game,” Javier says.

The strategy is to ‘solve’ a particular poker variation, having the computer play huge numbers of games against itself, adjusting its strategy each time. Sounds like a better plan than going back to school for another traditional degree.

If you want to lose your money to Javier, and you live outside of the United States, you can find him playing limit hold ’em on Just track him down before he gets his poker simulation up and running.