An open source framework for evaluating user input models
Python, Other · A framework for constructing keyboards and evaluating their performance with user interaction models. An example analysis where the… · More framework was used to find a keyboard layout that minimized interpretation errors when used with Swype input on touchscreens can be found here: http://sangaline.com/blog/optimizing_for_swype/
A breakthrough algorithm for particle identification in HEP
Other · Distinguishing between various particle species is an essential requirement in a vast range of analyses in nuclear and high energy physics.… · More Traditional methods require painstaking calibrations and tunings as well as somewhat arbitrary shape assumptions in order to extract particle yields. Pidrix is an algorithm and open source library that eliminates the months of effort that would have previously been required while simultaneously giving significantly more accurate results.
Python · The applicationbuilds a single animated SVG out of a sequence of distinct static SVG files. It has an experimental mode where it attempts to… · More find matching elements between frames an animate between them. This allows for smooth transitions between frames and also dramatically reduces the file sizes. When the SVGs were machine generated, say by matplotlib, this tends to work very well and produce stunning results.
An addictive card game that's like solitaire... but fun!
An extension for anonymizing and sharing Facebook posts
A tool for quickly evaluating fantasy football trades.
A mobile chat application with real-time translation.
My name's Evan and I'm a recovering physicist. I'm currently the CEO of Intoli. I spend most of my days writing code, analyzing data, designing algorithms, and talking to game developers. On the other days I like cooking, biking, and seeing how close I can get to alligators before I get too scared.
* I made the first statistical inference of the nuclear equation of state in high energy nuclear physics, a core goal of the field since its inception. This involved applying machine learning techniques to perform a stochastic optimization over the high-dimensional parameter space of a sophisticated nuclear collision model.
* I discovered a new set of statistical moments called strongly intensive cumulants. These quantities have the remarkable property that they can accurately measure important physics fluctuations in systems where the volume distribution is unknown.
* I developed an unsupervised machine learning algorithm for particle identification in high energy physics. Previous methods required many months of meticulous tuning and calibration in order to get reliable fits. The new method is totally automatic and outperforms traditional methods.
I'm mainly looking for game developers who want to increase their user engagement. Intoli is accepting new beta testers right now and we'll work one-on-one with our testers to help them improve their user retention and average game session length.