Resume | CV

Bradley J. Rhodes

Researcher, Inventor 510-735-7669

Experienced research scientist and technical lead with a demonstrated history tackling difficult technical problems that require multi-disciplinary solutions. Strong engineering professional skilled in Computer Science, Algorithms, Machine Learning, Artificial Intelligence (AI), Mobile and Wearable Computing, and UX/UI design. Publications in premier conferences and journals, and inventor on 45 issued patents.

  • Planning algorithms
  • Machine learning
  • Simulation & modeling
  • UX/UI design
  • Wearable computing
  • Ubiquitous computing
  • System architecture
  • Data analysis
  • Experimental design
  • Parallel processing (Flume & Apache Beam)
  • C++
  • Python

Loon LLC. (formerly X / Google X) Nov 2011 - Aug 2021

Tech Lead, Research Scientist

  • Designed and developed mission control and balloon simulation system, first as sole developer and later as team lead. The system was the primary tool for launch planning, flight tracking, and balloon recovery for the first three years of Project Loon, Google X's project to provide wireless internet to rural areas using high-altitude balloons.
  • Designed algorithms for steering balloons to a target by catching winds at different altitudes. Conceived and co-created algorithm that became the basis for all of Loon's long-term balloon planning algorithms.
  • Led team effort to simulate and visualize typical customer experience in different service regions and under different conditions. Results identified regions where service could be provided using significantly fewer balloons, leading to a major strategic shift to focus on tropical latitudes.
  • Conceived of static metrics to quickly compute ability to station keep over a service region, and developed tools to compute and visualize them. These metrics became the primary means of identifying potential service regions.
  • Conceived, designed and developed novel system for generating tunable synthetic wind noise for use in balloon simulation. The resulting system became a key component of Loon's simulation architecture.
  • Developed tools, metrics and data visualizations for evaluating wind forecast error across all historic flight data.
  • Took ownership of Loon's wind-mixer, a system for improving short-term wind forecasts using Gaussian processes. Developed changes that improved short-term forecast accuracy by 24%.
  • Designed and led development of architecture for integrating real-time storm data into Loon's simulation and planning architecture.
  • Ten issued patents.

Google X Apr 2011 - Oct 2011

Tech Lead, UX designer, Research Scientist

  • Designed, prototyped and evaluated UI and UX elements for what eventually became Google Glass.
  • Designed and ran UX study examining the physical limits of comfortable gaze angle, resulting in recommendations that were incorporated into the final design.
  • Six issued patents.

Ricoh eWriter Solutions Apr 2010 - Mar 2011

Senior E-Paper Interface Architect, Team Lead, Manager

Ricoh Innovations California Research Center 2001 - 2010

Senior Research Scientist


PhD, Media Arts and Sciences

Massachusetts Institute of Technology (MIT Media Lab)

Software Agents Group

Thesis: Just-In-Time Information Retrieval. Investigated the design of JITIR agents: software that suggests useful information based on a person's local context.
Thesis supervisor: Dr. Pattie Maes.

Master of Science, Media Arts and Sciences

Massachusetts Institute of Technology (MIT Media Lab)

Software Agents Group

Thesis: PHISH-Nets: Planning Heuristically In Situated Hybrid Networks. Presented a behavior-selection algorithm for autonomous 3D graphical robots in noisy environments.
Thesis supervisor: Dr. Pattie Maes.

Bachelor of Science, Electrical Eng. & Computer Science

Massachusetts Institute of Technology

Thesis: Doing the Right Thing Even Better. Evaluated and enhanced the Agent Network Architecture for behavior selection in autonomous robots.
Thesis Supervisors: Dr. Pattie Maes and Dr. Marvin Minsky.