Hello! Welcome to my portfolio
About
A brief introduction about me. My name's Max(Yinghao) Liu and I'm a senior majoring in computer science and data science at University of Michigan, Ann Arbor. I'm a programmer, graphics designer and photographer. I believe that tech brings people and organizations closer🧑‍ through innivation.
I'm not only passionate about artistic expression through tech, I also value intelligentual vitality in the process of coming up with interdisciplinary solutions 🔧 to real world problems. I'm intrigued by the design and implementation of intelligent systems, especially robotics🤖.
I’m actively searching for opportunities as a AI/Data infra engineer. I am a highly motivated and open-minded person. I boast strong academic and industrial experiences. Here's my Linkedin for a quick peek at my journey🌠.
Projects
Fetch Robot (2024)
Loaded the Fetch robot in simulation utilizing ROS Bridge and interactive hierarchy traversal to render custom URDF files. Implemented forward kinematics with LU decomposition routines to enable robot trajectory execution. Enabled posing robot end-effectors using inverse kinematics via gradient descent with Jacobian transpose. Integrated resolved-rate inverse kinematics with null space constraints to respective joint limits and Cyclic Coordinate Descent (CCD) inverse kinematics algorithm. Developed collision detection algorithm and both RRT- Connect and RRT-Star motion planner for robot navigation.
Minap (2023)
Code and documentation for Mi Nap sleep diary smartwatch app and related infrastructure, developed by the 2023 ITS intern cohort at the University of Michigan. This Fitbit / Apple Watch / Garmin smartwatch app allows clinical research participants to keep a sleep diary on their watch, with the data being securely transferred to a staging cloud storage service, and then to an Oracle PL/SQL database that sits behind a firewall.
ViewInsta Clone (2023)
Program an dockerized instagram clone that uses AWS RDS PostgreSQL database, AWS S3, and AWS CloudFront CDN for serving resources and handling uploads. Provision cloud infrastructure leveraging AWS CloudFormation. Configure load balancer and security groups on AWS EC2. Set up a Github actions CI/CD pipeline to update images on AWS ECR, register task definition and deploy application service on AWS ECS with AWS Fargate.
pSSID (2023)
pSSID engine is UMich ITS's latest effort to implement a full-stack web application for network analytics in collaboration with perfSONAR, an open source networking project. UMich is actively contributing to the project to make it possible to run on enterprise networks across research and higher education insitutions. I'm in charge of provisioning this engine on hardware Raspberry 4 probes. My script creates a daemonized process of continuously processing json config files generated by the application and passes the test information to perfSONAR.
ViewROS SAM Client (2023)
Devise an image segmentation ROS service for real-time segmentation mask generation with using OpenCV, PyTorch and NumPy. Design a GUI that allows customizable segmentation and manage multi-processes between GUI and ROS service. Separate local ROS service from the server hosting the segmentation model and provisioned server setup with Terraform.
ViewZotMeal (2021)
ZotMeal obtains a daily menu from the web and then displays it in a friendly user-interface. Both Brandywine and Anteatery’s menu is included with the most up-to-date information! Dining hall service hours can also be found in the app. Moreover, users are able to do a quick Google search on the food simply by clicking the menu. I'm in charge of writing the web scraper of the app's backend. The app is published and currently has 5000+ users.
ViewSnapShare (2021)
SnapShare uses machine learning to make donating your surplus food super easy. There are only 2 steps: Snap and Share. Snap a pic of your surplus food, and the machine learning model will automatically classify it and match a local food bank that needs it. This is the UVA Hackathon Social Empowerment Track Winner.
ViewPomodoro Cat (2021)
Working remotely can be challenging, especially amidst a worldwide pandemic. After analyzing the current applications that target productivity, we decide to integrate a fun game that allows users to keep their own cats while keeping track of their time using Pomodoro technique. This won the VTech Hackathon 2nd best overall.
ViewCodeAI (2021-2022)
This project integrates a natural language processor into a website using React.js, next.js and Firebase in order to build a learning platform for lower income elementary school students in Santa Ana, CA to learn AI literacy and possibly embark academic persuit in computer science. I worked on this project while serving as a research assistant at UCI Digital Learning Lab
View