Data Scientist, Cloud Architect
2013 - 2017 (over 3 years)
My primary responsibilities include designing, implementing, and leading production initiatives within the engineering team using the data collected by our wearable sensor...more products. Specifically, I've developed and lead the development of statistical analysis, machine learning models, visualizations, backend architectures, data logistics, and REST API's all of which have seen the light, and dark, of live production environments.
Enterprise Service Development with the goal of streamlining the intelligence gathering workflow. Responsibilities included integrating Virtual World software with other enterprise...more services as well as 3D modeling for data visualization.
Deploy a minimal, yet non-trivial Heroku app in minutes!
All the essential code you need to deploy a working Python Flask API to Heroku, with a PostgreSQL database, within minutes. It includes many features that a… · More REST API should have and illustrates how various components of backend code interact with each other. It can be used as a learning resource or the starting point for a hackathon / small project.
PhD Student · Graph-based (graphical) models are popular methods used in machine learning for visualizing and representing high-dimensional datasets.… · More One of the main problems, however, is estimating the parameters of the graphical model.
Gleason grading is typically done by doctors on a case by case basis. Since this process involves humans, grading is prone to several problems. First, the… · More task is extremely time-consuming and tedious as tissue samples must be examined in great detail. Second, when faced with many tissue samples a human grader may become fatigued and provide inconsistent scores. Third, it is impractical to expect humans to be the primary judge as medical data becomes increasingly digitized and stored in large quantities.
Accurately predicting the future behavior of a thread is an important component of intelligent thread schedulers designed for heterogeneous… · More multicore processors.In this paper, we focus specifically on the thread characterization task using performance counters.
My background is in computer science and machine learning. I did research in recurrent neural networks and other fields that became what we know now as deep learning.
I was the first engineer at the wearable tech startup Lumo BodyTech where I was involved in their entire tech stack from the firmware to cloud architecture. As the company grew I led their data science and cloud initiatives working with product, engineering, and marketing.
Currently, I'm the CTO and Co-Founder of FetchyFox working on building the worlds most sophisticated and intelligent e-commerce platform for airports.
I'm also a competitive climber (bouldering). I participate in local competitions and spend my non-startup time training.