Angshuman Guha has been a practitioner of applied machine learning for 24 years, spanning Microsoft, Google, Yandex Labs, Sears Holdings and a startup.
He has worked on handwriting recognition, optical character recognition, web search, image recognition, conversion prediction for online advertisements, ad click prediction based on text, machine learning infrastructure and other machine learning problems.
His main technical strength lies in deep learning (neural nets), but he has worked on all state-of-the-art machine learning technologies including random forests, boosting, bagging and support vector machines.
He chose to remain a hands-on engineer/scientist instead of going up the management ladder. He is passionate about open-ended hard computer science problems, with focus on, but not limited to, modern machine learning paradigms. He holds two dozen technical patents in the US and outside.
He has a BS in CS from Jadavpur University and an MS in CS from University of Iowa. He spent 3 years in the PhD program in CS at University of Wisconsin, Madison, but did not finish it.