James Bradley Wendt

About James Bradley Wendt

I am a Staff Software Engineer at Google DeepMind currently focusing on improving the capabilities of large generative models related to structural data understanding. Previously, I was at Google Research where I focused on methods for extracting and organizing structured data from templatic and unstructured content. I worked on a few different domains in this space, including private corpora such as email (e.g. extract the address and check-in/check-out times from a hotel confirmation email so that products such as Google Calendar, Assistant, etc. can surface the information back to the user on their respective platforms). I developed methods for Google Cloud\'s Document AI platform for processing form-like documents such as paystubs, bank statements, mortgage docs, etc.(e.g. extract fields such as the invoice date, amount due, line-items, etc.). And most recently I\'ve leveraged large language models to assist Google Search users when navigating small decision spaces (e.g. product comparisons when shopping online, comparing places of interest for travel or dining, etc.).I\'ve touched many different aspects of these problems, spanning modeling, data quality, data management, and infrastructure. Broadly speaking, I\'m interested in practical applications of ML.

Experience

1. Staff Software Engineer

   Google DeepMind

   May 2024 — Present

Education

• UCLA

  Ph.D., Computer Science

  2009 — 2015

• Pomona College

  B.A., Physics

  2005 — 2009

Skills

• Simulations
• Unix
• Artificial Intelligence
• Statistics
• Matlab
• Distributed Systems
• C++
• Mapreduce
• R
• Java
• Sql
• Linux
• Algorithms
• Software Development
• Python
• Teaching
• Latex
• Network Traffic Analysis
• Perl
• Computer Science
• Javascript
• Data Analysis
• Machine Learning
• Big Data
• Cloud Computing
• Research
• Software Engineering
• C
• Programming