Hi there, my name is Harry Liu! 

If you're here, you've either arrived through my resume, online application, or somehow guessed this URL and are just here by chance. No matter the reason you're here, I'd love to share with you my experience and portfolio in engineering and leadership thus far!

 

Go to different sections here:

Contact Info

Industry Experience

Project Experience

Leadership Experience

Industry Experience

 

Solutions Engineer - Zendrive

Oct. 2018 - Current

  • Closed highest ARR enterprise customer in our insurance segment, leading them through initial technical engagement, product and SDK discussions, SDK integration and debug, to validation and final product deployment.

  • Led and contributed to the pre-sales and post-sales technical process for customers in aggregate of 7-figures of ARR. 

  • Presented to, ran code reviews with, or supported 30+ customers (from small startups to enterprise companies) in the pre-sales, post-sales, integration, and support phase of working with Zendrive within my first 3 months as a Solutions Engineer.

  • Contributed to technical portion of RFP - led to a 7-fig ARR deal with a leading emergency services company.

  • Led efforts with Product, Engineering, and BD in refining our Insurance SDK integration documentation and validation process.

  • Isolated and identified a critical issue in our Android SDK with respect to drive tracking and gave feedback which was productized and iterated over by our SDK Engineering team.

  • Worked closely with a customer to debug sensor information, SDK functionality, and application logic with an influx of up to 10 unique support issues per day for this customer.

  • Created a full Cordova/Phonegap plugin from scratch using our existing SDK and wrote self-serve documentation on this plugin.

  • Integrated with customers that used native iOS, native Android, Cordova/Phonegap, React Native, or a third party application.

  • On-boarded as a Solutions Engineer within three weeks and quickly started providing value to customers from initial engagement to support.

  • Was the first Solutions Engineer to pass and officially be certified through our "SDK Certification Process".

  • Sat on final interview panel for two Solutions Engineers candidates and offered key feedback in the final decision making process.

Customer Success Manager - Zendrive

Jul. 2018 - Oct. 2018

  • Project management with key stakeholders (VP of BD, Head of Product, Insurance, Data, CS, etc) on next phase of Zendrive product offering.

  • Executed regular check-ups and engagements with 5 key customers within two months of starting my role.

  • Drafted and productized internal customer success milestones, handoff processes, and technical profiles to use as part of the CS lifecycle within Zendrive.

  • Presented on the Customer Success organization and responsibilities on my second day after starting my role.

  • Led, organized, and compiled results for a three week long internal test of our product for competitive analysis against three key competitors.

  • Organized a learning session for our internal SDK configuration and created an employee wiki page that educates on-boarding engineers on the specifics of this configuration.

Hardware Acceleration Design Engineer - Intel Corporation (Programmable Solutions Group)

Aug. 2017 - Current

  • Contributed through scripting and performance analysis on a low precision deep learning project on an Arria 10 FPGA for an internal Intel framework for neural network development.

    • Scripted in Python to automate compiling of .prototxt files for Caffe.

    • Scripted in Python to run automation tools for framework testing.

    • Scripted in Bash to run classification on 50,000 ImageNet validation images and compare accuracy with Caffe against internal Intel deep learning IP.

  • Performance analysis of several neural network topologies on Arria 10/Stratix 10 FPGA using Intel deep learning framework.

    • Analyzed SqueezeNet, AlexNet, ResNet, VGG, etc.​

    • Searched for optimal results based on vectoring, FPGA resources, bandwidth limitations, DSP usage, etc.

    • Guided several project goals by estimating throughput performance every week.

  • Verilog timing closure on an internal Intel IP that included several Verilog modules.

    • Improved setup and hold time by tens of nanoseconds.​

    • Able to improve slack time from negative to all positive.

    • Pipelined certain data pathways in order to meet timing without reducing overall clock speed.

  • Facilitated, ran, and documented PCIe & DMA driver validation for a shell module on Arria 10.

    • Modified and cut user and driver code to implement memory reads and writes to validate functionality through DMA.​

    • Configured FPGA and brought it up with SignalTap modifications in order to verify in hardware that the DDR calibrated successfully.

    • Documented driver validation process with a readme and file listing.

  • Led weekly seminars to the engineering team on key technologies that we use in development.

    • Went into the best practices of Verilog to ensure good and clean RTL development.​

    • Talked about latch inference, WYSIWYG, combinational vs. sequential blocks, etc.

College Grad Rotation Program - EMIF Applications Engineering & Data Center Solutions Engineering - Intel Corporation (Programmable Solutions Group)

Aug. 2016 - Aug. 2017

  • Implemented C++ API functionality for a variable amount of OpenCL devices to talk to Intel deep learning IP.

    • Wrote debug messages and expanded device functionality to run inference processes on multiple acceleration cards in parallel.​

  • Created VBA tool for high-end FPGA device selection based on memory interface and bandwidth requirements.

    • Learned VBA and used it to create a GUI based tool to decide on what FPGA device would fit the customer needs based on bandwidth (Gbps), DDR3/4, and efficiency​.

    • Implemented memory bandwidth formula dependent on bit width, operating frequency, data rate, and efficiency.

    • Documented tool in internal wiki page and produced a Youtube video for Intel FPGA documenting the tool.

  • Led competitive analysis on DDR parameterization tools between Xilinx and Intel and presented to team.

    • Used Quartus Memory Interface Parameterization tools and Xilinx IP EMIF Parameterization tools and compared the overall usability, support access, understandability, etc.​

    • Presented findings to the EMIF Applications Engineering team within Intel.

  • Initated and led an internal “Innovation” project for 10 recent college graduates that taught technical skills.

    • Wrote up a high-level proposal for the project and presented it to the manager of Applications Engineering and New User Experience Engineering.​

    • With the help of about 10 college graduates, created a top level module and several sub-modules with dual-port ROMs, RAM, and state machines for logic to instantiate a light sequencer on FPGA.

    • Created all modules from scratch.

    • Started a phase of the project that included integrated Arduino technology with onboard FPGAs for custom logic modules to accelerate the microcontroller.

Electrical Engineering Summer Intern (Circuit Design) - RefleXion Medical

Jun. 2015 - Sep. 2015

  • Worked with Principal Electrical Engineer to design a capacitive position feedback system for PoC.

    • Made a cylindrical concept design to serve as a sensing capacitor to measure capactive changes which give feedback for certain mechanical positions of the overall laser system.

  • Co-designed a front-end sensor board to detect gantry position and reduced rise time by twenty times.

    • Used PADS VX to design a PCB and ensure all connections were valid and functionality was implemented.​

    • Reduced signal rise time (from logic low to logic high) by twenty times compared to an old sensing system.

    • Utilized a metal tab on the gantry that passed through a sensor while rotating that triggered the sensor.

    • The sensor was then fed through a circuit (analog to digital conversion, full rectifier) to go to a logic high or low.

  • Circuit schematic design and simulation for subsystems with planning and expediting fabrication process.

  • Attended a Microsemi Libero SoC two-day course on the SmartFusion 2 Device.

Project Experience

 

Voice Activated Arduino Alarm Clock (click here for video demonstration)

Jun. 2015 - Oct. 2015

Thought up, created, and assembled a voice threshold sensitive alarm clock with the Arduino Uno platform from scratch. I used a couple of different components:

 - an RTC module (real-time clock) to keep track of the actual time, especially in the case of Arduino power shutdown (uses an onboard battery).

- a small adjustable gain microphone that had three pins (GND, VCC, and output pin to the Arduino).

- a Wave shield that required soldering of an SD card slot, potentiometer, and 3.5mm audio jack output.

- Arduino Uno platform.

- Resistor to ground. 

The purpose of the alarm is to prompt the user at a certain programmable time (i.e. 7:30 am). The alarm would then say "The time is now xx:xx, please shout after the beep". The user then has a 10 second window to shout above a certain threshold. If the shout is registered, the alarm is turned off and a verification is given to the user. If the shout isn't registered within 10 seconds, the alarm will sound a loud Klaxon. I prototyped the alarm on a breadboard and connected all the components together through wire and solder. I then took an example design from Arduino and coded a huge modification so that the input voltage levels from the microphone would correspond to threshold levels that trigger a certain response. The RTC would then allow for the alarm to trigger at a certain time consistently without needing power from the Arduino. A video documents this process below.

 

Autonomous Vehicle Design - UC Davis & NXP Cup (click here for 1st place run video)

Sep. 2015 - Jun. 2016

Senior design project that used the Freedom KL25Z development platform with a custom PCB that we designed to create an autonomous vehicle that could stay within two lines or track one line. The vehicle had two cameras to increase accuracy in tracking. The lines were fed through a 1-D buffer and we had to differentiate through grayscale whether the data in the shift register was white or black. You could adjust the sensitivity to lighting based on rotating the lens cap as well. We used a combination of histogram and PWM functions and used C programming with interrupts and UART serial communication. There were a few algorithms we implemented: boosting out of turns by sensing wheel position, compensating for low light based on a histogram, etc. We ended up winning 2nd place at the West Coast NXP Cup at UC Davis and won 1st place at the National NXP Cup in Austin.

 

Leadership Experience

 

Community Group Leader / Worship Band - Awakening Church

Jan. 2017 - Sep. 2018

Led group of 25+ in weekly bible study and music, organized community potluck events that 50+ attend quarterly, led discussion and analysis of topics and books in breakout groups of 12-14 men, played electric guitar monthly to bi-monthly for a group of 400-600 people.

Core Group Leader / Worship Team Leader / Officer - Davis Christian Fellowship

Mar. 2013 - Jun. 2016

Led weekly worship band practices and delegated musical roles, facilitated weekly bible studies for groups of 15, planned Welcome Week events (open mic and scavenger hunt) that over 250 people attended, managed a hospitality team of 8 people for weekly meetings.

Lights Circuit Lead - Solarri Controls

Jan. 2015 - Feb. 2016

Part of a bigger controls team that is designing a fully powered solar vehicle for competition. Designed and reviewed lights circuit diagrams with fellow team members. Decided on parts and specifications to use in vehicle while maintaining safety regulations. Coordinating with other team leaders on other subsytems in the vehicle's core control.

 

Contact Info