Kai Yu

Software Engineer / Machine Learning Engineer

I build games, application, and AI models.

About

My interest in computers started after playing countless video games, which sparked my curiosity about how computers and programming really worked. Since then, I've learned a lot during my time at Santa Rosa Junior College and as an undergrad at the University of California, Irvine. At UCI, I discovered a deep appreciation for Artificial Intelligence and its limitless potential.

I graduated in 2024 with a focus on Intelligent Systems and have built numerous machine learning models and softwares that I believe make a meaningful impact—not just by automating tasks, but by offering solutions where it is needed the most.

My most proud works are a deepfake audio classifier that will reduce the harm malicious deepfake AI will cause, electricity and settlement detection using satellite images for underdevloped Afrian villages, and hashing scraper that supports an agent application in preventing predatorial behavior in school and work enviornments.

Research

  1. 2024 — Present

    Topic Modeling and Analysis · UC Irvine

    Performed topic modeling using Latent Dirichlet Allocation (LDA) on student admission essays to extract topics by optizming coherence score in the model. Performed statistical analysis using the topics to correlate academic performance and probation.

    • Python
    • BerTopic
    • Octis
    • Statistics
  2. 2024 — Present

    ML Prediction on PFAS Contamination · UC Irvine

    Preprocessed Perfluoroalkyl and polyfluoroalkyl substances (PFAS) existing in groundwater for all 50 US States. Conducted statiscal analysis including descriptive statisctics, trend identification, and time-series analsyi on the dataset. Created visualizations including various plots and geospatial amps to help analyze data distribution and trends.

    • Python
    • PyTorch
    • Scipy

Experience

  1. 2023 — Present

    Software Engineer Intern · Netpropriate

    Build and maintain critical components used to construct Klaviyo’s frontend, across the whole product. Work closely with cross-functional teams, including developers, designers, and product managers, to implement and advocate for best practices in web accessibility.

    • Python
    • PostgreSQL
  2. 2023 - 2024

    STEM Insturctor · Dreams for Schools

    Co-instructed to 3-5th grade students with hands on engineering proejcts and basics

    • Engineering
    • Leadership
  3. 2021 — 2023

    Security Intern · County of Marin

    Assessed security risks and implemented security protocols in a team enviornment for the enterprise. Authored a 20 page security hadnbook, educating 2,400 county employees on cybersecurity threats. Conducted resaerch and created presentations for various security topics

    • SQL Injection
    • XSS
    • Social Engineering
  4. Oct — Dec 2020

    Web Developer · Sophie's Masasge

    Developed and styled website for Sophie's Massage Therapy.

    • React
    • JavaScript
    • HTML & CSS
    • AWS
  5. 2019 — 2020

    Student Tutor · Santa Rosa Junior College

    Computer Science and Math tutor helping students in 1 on 1 private sessions

    • C++
    • Java
    • Calculus

Projects

  1. Stable Diffusion Clone

    Stable Diffusin Clone

    Developed a Stable Diffusion clone in Python from scratch, utilizing diffusion techniques capable of generating images from text, images, and performing inpainting. The model incorporates CLIP embeddings and a variational autoencoder.

    • Python
    Stable Diffusion Clone Stable Diffusion Clone Stable Diffusion Clone
  2. deepfake audio

    Deepfake Audio Classifier

    Developed three deep leanring using different architecures to detect deepfake auido from real audio to prevent malicious AI generated content from deceiving unsuspecting victims.

    • LSTM
    • Tensorflow
    • Transformers
    • Wav2Vec2
    deepfake audio deepfake audio deepfake audio
  3. settlement detection

    UNet3+ Settlement Detection

    Implemented UNet3+ model for advanced electricity detection using satellite data on African settlements. The dataset used comes from the IEEE GRSS 2021 ESD dataset composing of 98 tiles representing a 64KM area.

    • PyTorch
    • WandaDB
    • PyTorch Lightning
    settlement detection settlement detection settlement detection
  4. minecraft

    Minecraft in C++

    Created a Minecraft clone from scratch using OpenGL, featuring custom rendering, texturing, and player physics. The game allows players to build, destroy, and move around, closely mimicking the mechanics of the original game.

    • C++
    • OpenGL
    minecraft minecraft minecraft
  5. cysafe

    Cysafe

    HackDavis 2023 winner, this project enhances biker safety by integrating LED blinker lights, a rear-view vehicle detection and alert system, and an emergency contact messaging system directly into the helmet.

    • Python
    • OpenCV
    • Arduino
    • Raspberry Pi
    cysafe cysafe cysafe
view all projects