Background
Peter Groves Headshot 2019 I’m a Machine Learning Engineer at the National Center for Supercomputing Applications on the Visual Analytics team. As the only backend engineer on a team that delivers full stack apps for medical researchers, I have significant expertise in API development, devops, cloud engineering (GCP, AWS, Openstack), Docker, Postgresql, and various pipeline technologies.

I first began applying machine learning techniques to real-world data as a graduate student in 1999 including industry projects with Allstate and Caterpillar. My research group spun out a startup called Riverglass where I continued consulting-style projects with BP (and later Deere) while also being the technical liason on the sales team for two years.

Data science projects that never advanced beyond a proof-of-concept became boring, however, so I transitoned to engineering roles where I designed, implemented, and operated machine learning components as part of production software systems. For seven years I did this as a contractor to startups in the Chicago area and while also working on my own startup projects.

During my time at Groupon, I worked on a team that delivered product recommendations through an API to both web and mobile applications. That service required continuously updated machine learning models, served thousands of requests per minute, and had over 99% uptime.

Now that I have returned to academic research with the National Center for Supercomputing Applications (NCSA), I work on production machine learning tools that can handle complex real-world data. Most of my projects involved the analysis and visualization of medical data for researchers. Recently, I have focused on structuring, storing, and accessing complex data sets so downstream machine learning tasks can be carried out without algorithm-specific pre-processing.

I currently lead the development of the VISA team’s OmiX platform and Pixsure, a tool to annotate scientific images to train ML models to perform segmentation (finding the silhouette of each detected object).

Quick Contact
[email protected]
linkedin.com/in/petergroves
twitter.com/designbyrobots
github.com/pgroves
Sitemap

Full Resume (with pdf)

More Social Media Links

Consulting info, although I don't have much time for it these days