Taylor Brown Email & Phone Number

Texas, United States

Irvine, CA, US

Started and leads the computer vision and deep learning team.Key projects:* Architected and delivered the solution for first floor height that FEMA uses for risk rating 2.0. Developed patented method for unconstrained photogrammetry.* Designed and oversaw implementation of appraisal collateral review checks based on imagery.* Oversaw multiple imagery and.

Irvine, CA, US

• Principal technical team member on several machine learning research projects for the insurance and real estate markets. Key Contributions:
• Led team to develop a number of deep learning imagery models using both overhead and natural imagery. Responsible for setting plan and timelines, managing performance to plan, and communicating results and strategy to.
• Developed deep learning models to classify building footprints. Constructed entire geo- and image processing pipeline to classify footprints from aerial imagery. GDAL/python/keras/tensorflow
• Identified high fire risk homes using NLP techniques on MLS data. Used word/doc2vec
• Created models to predict key attributes of every property in the USA. Implemented in production for a significant commercial product. Used random forest and xgboost.

Charlottesville, Virginia, US

• Performed machine learning and software development across a wide variety of problems.Key contributions:
• Researched and analyzed Deep Convolutional Models. Researched and developed state-of-the-art geospatial predictions using deep neural networks. Developed and promoted first in-house expertise in deep learning.
• Led team of 4 to develop new modeling techniques. The approach merged relational ensembles such as gradient boosting with relational graph data. Implemented in scala.
• Designed and implemented NoSQL GPS track system. System reconstructed vehicle tracks from sparse/noisy GPS logs using a specially modified A*.
• Contributed to a method to store the data in a distributed, geographic database, now known as geomesa. http://www.geomesa.org/
• Researched and analyzed Random Forest model interpretation. Using dimensionality reduction methods such as spectral clustering, developed method to extract underlying structure in the model, providing end-users with.

Arlington, VA, US

• Development support for a research and design initiative to change the way Boeing reused parts and assemblies.Designed, implemented and improved several tools to increase productivity and streamline business processes..
• Designed and implemented web application to facilitate configuration management and establish business processes, while increasing customer satisfaction. Implemented in C#.
• Designed and implemented heuristic test log verification tool for verifying different and unknown log formats during testing with C# and regular expressions. Saved time and potential human error in weekly log check.
• Maintained and improved python workflow tools.

Arlington, VA, US

• Improved work-flow tool used by engineers on 787 production line. Key Contributions:
• Worked with a team to improve a work flow system using.NET. Refactored key components to improve ease of maintenance and performance issues.
• Designed and developed a tool for generating schedules based on personnel constraints and job priorities using C#.
• Designed, developed and implemented a tool for generating a visual time based display of schedules in Visio using C#. Allowed team of 40 engineers to generate visuals automatically instead of by hand.

Norman, Oklahoma, US

• Responsible for IT department of the School of Geology.Key Contributions:
• Led team to develop new web based record system for School of Geology using an MVC framework in PHP and a MYSQL database on a Linux server. Established testing guidelines, implemented source control and code reviews.
• Managed team of 6 to handle technical support for the college.

Master Of Science, Computer Science

Bachelor Of Science, Industrial Engineering