index.html

<!DOCTYPE html>
<html lang="en">

<head>
  <meta http-equiv="X-UA-Compatible" content="IE=edge,chrome=1" />
  <meta charset="UTF-8" />
  <title>Julian Forsyth | Researcher & Engineer </title>
  <meta name="viewport" content="width=device-width" />
  <link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.0.0-beta3/css/all.min.css" />
  <link rel="stylesheet" media="all" href="assets/css/styles.css" />
  <link href="https://fonts.googleapis.com/css?family=Roboto+Mono&display=swap" media="all" rel="stylesheet" />
  <script src="assets/js/scripts.js" defer></script>
</head>

<body>
  <header>
    <h1>
      Julian Forsyth<br />
      --- <br />
      Researcher, Engineer
    </h1>
  </header>
  <header id="links">
    <h1>
      <a class="icon" href="mailto:julianforsyth@hotmail.com" aria-label="Email">
        <i class="fa-solid fa-envelope"></i>
      </a>
      <a class="icon" target="_blank" href="https://github.com/j-forsyth">
        <i class="fab fa-github"></i>
      </a>
      <a class="icon" target="_blank" href="https://www.linkedin.com/in/julianforsyth/">
        <i class="fab fa-linkedin"></i>
      </a>
      <a class="icon" target="_blank" href="assets/Resume 2025.pdf">
        <i class="fa-solid fa-file-pdf"></i>
      </a>
    </h1>


  </header>



  <br />
  <div id="container1-wrapper">
    <div id="container1">
      <!-- <h2>About me</h2> -->
      <section id="about-me">
        <p>
          I’m an AI researcher and engineer with experience discovering 
          universal concepts in vision and language models,
          steering robotic wheelchairs safely to their destination,
          and forecasting energy demand from buildings to regions.
        </p>
        <p>
          I care about making AI systems that are safe and understandable —
          especially in domains where reliability is non-negotiable and mistakes
          have real human consequences.
        </p>
        <p>
          In my spare time I try to get outside as much as possible, and watch great and/or terrible movies.
        </p>
      </section>
    </div>
    <div id="profile-image">
      <img src="assets/img/j.jpg" width="240" alt="Portrait of Julian" decoding="async" fetchpriority="high" />
    </div>
  </div>

  <div id="work" , class="container">

    <!-- <section id="video-demo" aria-label="In-view video demo"> -->


    <h2>Latest Work</h2>
    <a href="http://yorkucvil.github.io/UniversalSAE/demo">
      <section class="media-card" role="group" aria-labelledby="mediaTitle">
        <div>
          <h3 id="mediaTitle">Interactive Concept Explorer</h3>
          <p>A demo for our Universal Sparse Autoencoder project at ICML 2025! Made in collaboration with Thomas Fel.
          </p>
        </div>
        <div>
          <video id="demoVideo" preload="auto" muted loop playsinline webkit-playsinline
            controlslist="nodownload noplaybackrate">
            <!-- Provide both formats if you have them -->
            <source src="assets/img/demo.webm" type="video/webm" />
            <source src="assets/img/demo.mp4" type="video/mp4" />
            Sorry, your browser doesn’t support embedded videos.
          </video>
        </div>
      </section>
    </a>

    <a href="http://yorkucvil.github.io/UniversalSAE">
      <section class="media-card" role="group" aria-labelledby="mediaTitle">
        <div>
          <div class="slideshow" data-io-scale>
            <div class="slide-track">
              <img src="assets/img/team.jpeg" alt="Universal Sparse Autoencoder Screenshot" />
              <img src="assets/img/USAE.png" alt="Universal SAE Additional Image" />
            </div>
          </div>
        </div>
        <div>
          <h3 id="mediaTitle">Presenting at ICML 2025!</h3>
          <p>We presented our paper "Universal Sparse Autoencoders:
            Interpretable Cross-Model Concept Alignment" at ICML 2025 in Vancouver.
          </p>
        </div>
      </section>
    </a>


    <h2>Publications</h2>
    <section class="section">
      <article>
        <h3>
          "Universal Sparse Autoencoders: Interpretable Cross-Model Concept Alignment"
        </h3>
        <p class="text-small">
          2025 International Conference on Machine Learning (ICML)<br />
          <a href="https://doi.org/10.48550/arXiv.2502.03714">Access paper here</a>
        </p>
      </article>
      <hr />
      <article>
        <h3>
          "Robot Wheelchair Convoys for Assistive Human Transportation"
        </h3>
        <p class="text-small">
          Proceedings of the 2024 IEEE International Conference on Smart Mobility<br />
          <a href="https://doi.org/10.1109/SM63044.2024.10733535">Access paper here</a>
        </p>
      </article>
    </section>
    <h2>Work</h2>
    <section class="section">
      <article>
        <h3>
          Research Assistant,
          <a href="https://yorkucvil.github.io" target="_blank">Computer Vision & Imaging Lab</a>
        </h3>
        <p class="text-small">
          @ York University, 2024 - present<br />
          - Collaborated with an inter-institutional team to conduct cutting-edge research focused on
          interpretability in deep neural networks<br />
          - Designed and implemented an innovative training framework<br />
          - Developed statistical and visual tools to communicate our findings and reveal model insights<br />
          - Contributed as second author to a paper for ICML 2025 (International Conference on Machine Learning)
        </p>
      </article>
      <hr />
      <article>
        <h3>
          Software Engineer & Researcher,
          <a href="https://www.elderlab.yorku.ca/" target="_blank">Elder Laboratory</a>
        </h3>
        <p class="text-small">
          <strong>NSERC Undergraduate Student Research Award</strong><br />
          @ York University, 2023<br />
          - Developed motion control and obstacle avoidance protocols for an
          autonomous wheelchair robot.<br />
          - Collaborated with robotics team made up of engineers and developers under Dr.
          James Elder.
        </p>
      </article>
      <hr />
      <article>
        <h3>Teaching Assistant</h3>
        <p class="text-small">
          @ York University, 2022 - Present<br />
          - Guided students in Java coding, Object-Oriented Programming
          concepts and data structures.<br />
          - Helped students debug their code and work through challenges.
        </p>
      </article>
    </section>

    <h2>Education</h2>
    <section class="section">
      <article>
        <h3>MSc in Applied Computing</h3>
        <p class="text-small">@ University of Toronto, Sept 2025 – Dec 2026 (expected)</p>
      </article>
      <hr />
      <article>
        <h3>B.A. Honours in Computer Science</h3>
        <p class="text-small">@ York University, 2020 - 2024, GPA: 3.9 <br>
          Graduated First Class with Distinction</p>
      </article>
      <hr />
      <article>
        <h3>B.A. in Philosophy</h3>
        <p class="text-small">
          @ Concordia University, 2017 - 2019, GPA: 3.8
        </p>
      </article>
    </section>

    <h2>Projects</h2>
    <section class="section">
      <article>
        <h3>Short Term Load Forecasting with GANs</h3>
        <p class="text-small">
          @ York University, 2024<br />
          - Exploring Generative Adversarial Networks as a tool for predicting energy usage
          on residential power grids. Supervised by Dr. Michael Jenkin.
        </p>
      </article>
      <hr />
      <article>
        <h3>Syntax Writing Assistance Tool</h3>
        <p class="text-small">
          @ York University, 2023<br />
          - Created a tool to improve word choice in text using GPT-2 and
          BERT.
        </p>
      </article>
      <hr />
      <article>
        <h3>E-Commerce Platform</h3>
        <p class="text-small">
          @ York University, 2023<br />
          - Built a model marketplace with Java Spring, SQL backend, and React
          frontend, deployed on AWS.
        </p>
      </article>
    </section>

  </div>
</body>

</html>