Designed and built out an implementation of auto categorizing financial transactions. Developed a global vectorizer with a local classifier and clusterer to allow for bulk...more individual categorization. Spec'd out the pipeline to scale this. Front and backend feature work to support various features and a dynamic reporting engine.
Developed a clustering algorithm to identify neighborhood pockets across North America. Designed a distance function to assess neighborhood similarity. Built a model to predict...more amenity saturation. Wireframed, prototyped and designed product spec. Implemented front end in React+Reflux and backend in Python+Pyramids.
Built a realtime TTS model with multiple speakers and prosody. Built models to predict engagement, aesthetics, and shareability with users on photos. Built a pipeline to scale to billions of inferences on photos. Pioneered and prototyped various financial algorithms. Researched, designed and implemented a location analytics recommender tool.
One thing you quickly notice about Zak is his ability to brainstorm huge concepts and then boil them down to an appropriate level. Beyond this, he learns quickly and has top-notch hacking skills. When he sinks his teeth into an idea, he can run with it and hack together a prototype with surprising speed. Zak is friendly, funny, easy to get along with, and an asset to any team.