about

i'm a machine learning / computer vision engineer. i received a PhD from UCSD where I was advised by serge belongie. most of my graduate work was funded by an NSF IGERT traineeship and a 2010 google fellowship, and revolved around weakly supervised learning and its applications to object detection, recognition and tracking; one of my algorithms is included in OpenCV 3.0. after finishing my PhD, I started a company called anchovi labs which was acquired by dropbox; i worked at dropbox for 2.5 years. i spent the next 3 years at orbital insight, where i developed computer vision algorithms to analyze satellite imagery. i currently work on medical imaging applications at google. oh, and if you haven't noticed already, i have no respect for capital letters.

in my spare time i play guitar, read/watch sci-fi, and occasionally blog/tweet.

blog posts

• 2023-08-25 - sci-fi books
• 2018-04-27 - weight decay vs L2 regularization
• 2018-04-23 - practical tips (life, machine learning, and everything)
• 2017-10-01 - precision and recall
• 2017-08-25 - moving my website and blog to github
• 2016-03-27 - multiprocessing and seeded RNGs
• 2015-01-03 - deep net highlights from 2014
• 2014-11-29 - template matching basics
• 2014-05-24 - learning low-level vision feautres in ~10 lines of code
• 2014-04-20 - convolutional learnings: things i learned by implementing convolutional neural nets
• 2013-08-12 - dear zappos: here’s a way you can significantly improve user experience
• 2013-07-28 - when hogs py
• 2013-05-13 - harry potter and the curse of dimensionality