Michael Liu

Michael Liu

About Me

My name is Michael Liu, I am an undergraduate studying computer science at Carnegie Mellon University. I am particularly interested in computer systems, competitive programming, and building autonomous systems. Outside of computer science, I fence (competitively for about a decade now), enjoy spending time with friends and family, and embark on projects that challenge me.

Currently, I am a research assistant in the robotics department at Carnegie Mellon University, a web developer for ScottyLabs as part of the tech team for CMUEats.com, and a teaching assistant for CMU's Introduction to Computer Science course. Thanks for checking out my website!

Selected Projects

These are some of my projects and research so far.

  • OncoReconcile AI (A web app for normalizing disease, gene, and variant names and scoring reconciliation confidence)

    • Python
    • FastAPI
    • React
    • Tailwind CSS
    • Oncology

    Platform for normalizing disease, gene, and variant names with curated catalogs and evidence from APIs (ClinVar, CIViC, MyVariant), plus human-governed review for difficult cases. Also offers patient journey analytics and FHIR/OMOP JSON export prototypes.

  • Exploration and Implementation of Web Page Ranking Algorithms: PageRank, HITS, and Personalized PageRank

    • Python
    • linear algebra
    • PageRank
    • HITS
    • Personalized PageRank

    Exploration of the linear algebra concepts behind the PageRank and HITS algorithm, their implementation in Python, and further extensions involving Personalized PageRank.

  • Autonomous Vehicle Control: Imitation Learning and Soft Actor-Critic Integration for Enhanced Performance

    • imitation learning
    • soft actor–critic
    • reinforcement learning
    • Python
    • autonomous vehicles

    Research project integrating imitation learning and soft actor–critic methods for improved autonomous vehicle control and performance.

  • LeapASL (A realtime ASL gesture recognition and translation interface based on machine learning)

    • Java
    • Ultraleap SDK
    • Swing
    • neural network
    • SVM

    Independently built a program using Leap Motion to capture 31,000+ ASL letters, trained a sequential neural network and SVM, and integrated the models with a Java Swing GUI to construct and correct text in real time.

  • Flaurapedia (An android application for plant identification and blogging)

    • Java
    • XML
    • Android Studio
    • TensorFlow Lite
    • Firebase

    Independently developed a mobile app built with Android Studio (Java/XML), TensorFlow Lite, and Google Firebase featuring user authentication, plant identification via image upload or capture, and community post sharing/browsing.

  • VolunTEEN (A Flutter-based volunteering platform with Firebase authentication and Google Maps integration)

    • Dart
    • Flutter
    • Firebase
    • Google Maps
    • authentication

    An app using Dart and Flutter with user authentication, location-based volunteer matching, and event organization to connect event hosts with local volunteers.

  • Bitmap-to-Sketch Pen Plotter Using Raspberry Pi, Python Modules, and Servo Control

    • Raspberry Pi
    • Python
    • hardware

    Constructed a pen plotter machine that uses a Raspberry Pi, servo motors, integrated Python modules, and simple household items to convert bitmap images into physical sketches.

  • Old Games I Made for Fun!

    • Unity
    • C#

    A collection of old 3D platformer games in high school that I made for fun using Unity and C# in my free time. Likely buggy and unoptimized but I had fun making them and learned a lot about game development. Check them out and play them in your browser at my Itch.io!

Recent News

  1. started computer science at carnegie mellon university.