Jake Garrison

I'm Jake. I like designing, making, breaking and solving and fixing things

I studied Electrical Engineering at the University of Washington as an undergraduate and specialized in machine learning, computer vision, and signal processing for graduate school. I was born and raised in Spokane, WA and am now based in Seattle, WA. In my free time, I build various (mostly) open source projects from software to embedded devices to analog music gear.

Academic Epochs

High School

I designed and built my own electric vehicle by converting a seized Volkswagen car to run fully off of batteries. I also designed and built custom analog/digital audio hardware. It still drives.

College

I continued working with electric cars as one of the leads for the UW EcoCar team. This led to an internship at Tesla where production firmware code for the Model S Insane Mode and Model X Falcon Wing doors and Autopilot. Following this, I focused on autonomous driving and in-vehicle display research via EcoCAR. During this time I continued to build audio hardware and software. I briefly worked at Verellen Amplifiers, researched various sound synthesis techniques, and also built rtunes, a cross-platform music discovery app based on crowdsourced data from Reddit and AI. I also began contributing to hardware and software related to cryptocurrency mining and 3D printing.

Graduate School

I was a researcher at the UbiComp lab. Specifically, I worked on health screening/sensing mobile applications based on audio, video and other sensor input. My thesis was on the topic of smartphone sound based spirometry powered by AI. I explored several other AI, machine learning powered solutions as part of Dedsimple, a company I co-founded. Additionally, I helped develop Zuru, an AI design assistant service offered by Haiku Deck.

Resume

Education

  • University of Washington
    2011-2018

    B.S. and M.S. in Electrical Engineering

    I enrolled in UW as an undergraduate and dabbled in Digital art, Mechanical Engineering, Applied Physics, Design, Music, Computer Science and settled with Electrical Engineering. I became fascinated with signal processing due to its wide realm of mediums (sound, image, video, data) and made that my focus. During my masters, I did research in the areas of mobile health, computer vision, sound, natural language, automotive and machine learning in the UbiComp lab.

Work Experience

  • UbiComp Lab
    2016-Present

    Website

    The ubiquitous computing (UbiComp) research lab, led by Prof. Shwetak Patel, focuses on many areas of ubiquitous computing including novel user interface technology, energy sensing, low-power sensing, health sensing, and activity recognition by applying expertise in sensing, signal processing, embedded systems, circuits, and human-computer interaction. Close collaborations with Microsoft Research, Oculus Research, Intel, Google, and many other companies.

    • Using signal processing and machine learning to measure lung health and screen for obstructive lung disease on a mobile phone
    • Designing a mechanical pulmonary system
    • Researching methods to measure blood pressure via mobile phone camera
    • Collect data from patients and students
    • Mentor undergrad students
  • Puppy.ai
    2015-Present

    Website

    An iOS dog breed detection app assisted by on-device deep learning powered by Tensorflow. puppy.ai is available in the app store. Worked on a small, effective international team of three.

    • Trained Tensorflow model and build mobile engine
    • Designed iOS app (App Store link) and web app
  • rtunes.xyz
    2016-Present

    Web App

    Created open source music player that generates hundreds of playlists and stations based on crowed sourced data from Reddit. Select playlists that answer questions such as "What's your favorite song in a language you don't speak", or get a live feed of music as people around the world post it.

    • Open source and used by thousands of people
    • Built web and mobile apps with React.js
    • Distributed desktop app powered by Electron
  • Haiku Deck
    Summer 2016

    Demo

    A presentation iOS and Web app Seattle startup used by millions. It features Zuru, an AI platform for presentation automation. I contributed to research and development for the Zuru service and its assistant features that generate presentations from external content such as wikipedia or outlines.

    • Startup experience
    • Contracted to write production ready code
    • Web development mostly with React and Node
  • Driver Awareness System
    2015-2016

    Video Demo Interactive Demo

    Founded and lead a student team designing a full-fledged driver awareness system to monitor a driver's attention using metrics such as distraction, sleep, phone usage and happiness. Metrics are based on realtime analysis of the drivers face and can be visualized on our UI. This concept won Innovative Use of Data (GE) and Best Travel Hack (Concur) at its original unveiling at DubHacks, then 3rd place Consumer Appeal at the national EcoCAR event. Entered in Amazon Catalyst for funding.

    • Microsoft Kinect embedded vision hardware
    • Vision detection algorithms (C/C++, OpenCV, Matlab)
    • Frontend data visualization (D3 and Javascript)
    • Backend data filtering and processing (Python, ZMQ)
  • Urban Parking
    2016

    City of Seattle Parking App project sponsored by the Seattle Department of Energy to forecast available parking and route drivers to open spots. Forecasting is based on 45 million historical transactions. Future collaboration with the Google Sidewalks Project. Submitted to UW CoMotion for further funding.

    • Custom machine learning algorithm
    • Fullstack web and Android app development
    • Several Presentations and project pitches to investors and educators
  • Tesla Integration Team
    Summer 2015

    Model X

    Systems Integration intern on a small multidisciplinary team focused solely on the development of the Model X falcon doors and sensing.

    • Wrote production code
    • Presented and submitted reports to engineering VPs including Elon
    • Contributed to Model X: controls, firmware, sensing, design, validation and testing
    • Created custom data collection and visualization tools
    • Built custom DC motor dynamometer and shared the design with SpaceX engineers
    • Very fast paced
    • Set up and documented Phase Space 3d motion modeling system
  • Advanced Driver Assist Lead
    2015-2016

    EcoCAR 3 Youtube Channel

    Lead the autopilot research team for EcoCAR 3 which aims to design and integrate advanced autopilot features on a 2016 Camaro. Our system is capable of realtime object detection including sign, lane, vehicle and pedestrian detection. Received Second Place for Driver Assist technology (tie)

    • Project Management and leadership
    • Developed realtime vision applications based on the YOLO convolutional neural network framework and deployed on the NVIDIA Jetson TX1 Embedded Systems Module
    • Embedded Freescale S32V hardware for CAN signal and vision processing
    • Employed sensors such as stereo video, ultrasonic, gps and radar
  • Bankroll Bitcoin Miner
    Summer 2014

    One of three working on consumer bitcoin startup project. Contributed to PCB, firmware development, schematic and plastic enclosure. Worked with custom ASIC mining chip

  • Tesla Power Electronics
    Summer 2014

    Model S

    Power Electronics intern focused on the high voltage electric system for the Model S P85D Insane mode

    • Injection Molded and machined part design and tooling
    • 3d printing
    • Testing and validation on existing parts (temperature and current cycling)
    • R+D for future technology (specifically high voltage fuses, contactors and charging plugs)
    • Creating and implementing professional CAD, drawings, schematics, and PCB designs

  • Electrical Lead
    2013-2015

    EcoCAR 2

    Involved in the competition for all three years and became electrical lead . We were awarded first place electrical team for multiple years

    • Lithium ion battery pack design
    • High voltage training, fusing, simulation, harnesses
    • Modeled high and low voltage systems in Simulink, Matlab and NX CAD
    • Created and implemented professional schematics and wire harnesses
    • Lots of leadership and project management
    • Collaboration with multidisciplinary students, faculty and industry
  • Focused Ion Beam
    2013-2014

    Research Assistant for Professor Bruce Darling, UW Electrical Engineering.

    • Simulated circuit boards in Multisim and Labview
    • Designed PCBs with Ultiboard
    • Worked with one of the few remaining FIB machines
  • Verellen Amplifiers
    Summer 2013

    Website

    Worked in a Fremont, Seattle startup that builds custom boutique tube amplifiers and analog distortion pedals.

    • Board populating and soldering
    • Amplifier schematic design
  • Custom Music Equipment
    2011-2014

    Designed, built and sold guitar pedals and analog audio FX for local bands and friends

    • Learned analog electronics fundamentals
    • Became familiar with PCB design, entrepreneurship and part sourcing
    • Lots of troubleshooting and breadboard work
  • Electric Car Conversion
    2010-2012

    Website

    I designed, funded and converted a gas powered Volkswagen GTI to fully electric while in highschool. Awarded Washington Society of Engineering scholarship upon completion

    • Designed and hand built the high voltage motor controller
    • Sourced all of the new components , sold the old
    • Operated with a small, personal budget, funded by my job at the local bakery
    • Built custom metal components and wiring harnesses

Projects

Machine Learning

Wikipedia Summarizer
Urban Parking App
OpenAI Gym Experiments

Vision

phone.ai
puppy.ai
Reverse Image Search
Cartoon Filter GUI
Matlab Image Processing Tools
Gaze Tracking

Sound

rtunes.xyz
C Instrument Tuner
Online Music Scraper
Audio Seperate
Supercollider Compositions

Automotive

Driver Awareness System
Dashcam Vehicle Distance
Tesla Model X Falcon Doors
EcoCAR 3
Autonomous Vehicle Simulator
Lane Tracking
Battery Pack Design
Tesla Model S Duel Motor
EcoCAR 2
Electric Car Conversion
Homemade 500 Amp DC Motor Controller

Audio Hardware

Realtime Pitch Shifter
Verellen Amplifiers
Homemade Analog Audio
Arduino Synthesizer
RC Car Composer

Fabrication

Gameboy Color CAD

Papers

Smartphone AI
Spiro Sound
Deep Sound Classification
Urban Parking AI
FPGA Pitch Shifter
Cartoon Image Filter
Driver Assist Writup

Presentations

Pitch Shift
Lane Detection
Electric Car Systems

Posters

Phone AI
Spiro Sound
Urban Parking
Driver Assist Consumer Appeal