Home Projects Resume Contact

    Kevin He

    contact@kevinh.dev
    +64-451-781-138
    kevinh.dev
    kevinh-e

    Education

    Education

    Education

    University of New South Wales
    Sydney, NSW
    BA in Computer ScienceFeb. 2023 – Oct. 2026
    Normanhurst Boys High School
    Sydney, NSW
    Higher School CertificateFeb. 2017 – Nov. 2022

    Technologies

    Languages
    Python
    Java
    JavaScript/TypeScript
    C
    SQL (Postgres)
    Bash
    Lua
    Frameworks and Libraries
    React
    Next.js
    Node.js
    PyTorch
    TailwindCSS
    TanStack
    MongoDB
    Developer Tools
    Git
    Docker
    Linux
    NeoVim
    Markdown
    Figma
    Vercel
    Copilot

    Projects

    Deep NN Chess Engine
    2025
    PyTorch
    Tensorboard
    Google Colab
    • Designed a neural network–based chess engine using self-play and MCTS inspired by AlphaZero
    • Accelerated training by leveraging multiprocessing, batched inference, mixed precision, and layer pruning to save training time by 1700% versus a naive single-CPU loop
    • Streamlined a self-play data generation pipeline alongside supervised learning in Python to bootstrap engine performance to ≈ 2000 ELO within only 10 hours
    • Achieved a competitive ELO on custom test suites against Stockfish-level tactics while training cost was free
    TasteBuds
    2025
    Next.js
    React
    Socket.io
    Node.js
    TailwindCSS
    Git
    Shadcn/UI
    • Developed a gamified cuisine recommendation application with real-time multi-user interaction via Socket.io
    • Leveraged Shadcn/UI and other component libraries to design a modern, responsive frontend
    • Optimised developer workflows and productivity using GitHub to deploy a working MVP within 3 days
    • Led a 4-person development team to a 2nd‑place finish at the 2025 DevSoc Hackabyte Hackathon
    COVID19 Classifier
    2025
    PyTorch
    Python3
    Ray
    Tensorboard
    SKLearn
    • Self‑taught machine learning concepts from scratch to modern ResNets in PyTorch over 7 days
    • Developed a custom modular 56‑layer ResNet trained on 33 920 chest X‑ray images to identify COVID‑19, Lung Opacity, and Viral Pneumonia
    • Integrated Ray Tune and TensorBoard to visualise training metrics and auto‑optimize hyperparameters, filtering outlier runs
    • Achieved 92.93% classification accuracy on CIFAR‑10, matching top models from the original ResNet paper

    Experience

    Full‑Stack Developer
    Google AI in Action Global Event
    Sydney, NSW
    May 2025 – Jul 2025
    • Spearheaded a 3‑person design team to launch a B2B SaaS prototype using React, Node.js, and MongoDB
    • Implemented industry‑standard security with Clerk, Google API integrations, and serverless CI/CD pipelines via GitLab CI → Google Cloud Run & Firebase
    • Managed end‑to‑end project planning, integration testing, build standardization, and deployment of containerised services on Google Cloud using Docker
    Private Tutor
    Apr 2023 – Present
    • Provided one‑on‑one tutoring in Mathematics and Physics for HSC‑level students
    • Tailored lessons to suit varying learning styles and behaviours, achieving a 90% success rate in target grades
    • Created custom practice materials and simplified complex concepts to effectively teach Extension 1 Mathematics

    Extracurricular

    • Participated in the Titanic – Machine Learning from Disaster Kaggle competition
    • Participated in DevSoc Hackabyte Hackathon, achieved 2nd place
    • Built custom PCs; highly experienced in hardware assembly and troubleshooting
    • Earned an AMEB Certificate of Performance for Piano (Piano Comprehensive Solo Grade 8)
    Last updated in June 2025