resume

Education

  • 2022-2026
    UC Berkeley
    B.S. in Electrical Engineering and Computer Science
    • Clubs
      • Accel Scholars
      • Machine Learning @ Berkeley
      • Society of Women Engineers
      • Association of Women in EECS
    • Coursework
      • Machine Learning, Operating Systems, Computer Graphics, Algorithms, Computer Security, Data Structures, Linear Optimization Models, Discrete Math and Probability, Structure and Interpretation of Computer Programs, Designing Information Devices and Systems I and II

Experience

  • Aug 2025 - Present
    Google
    Software Engineering Intern
    • Integrating JAX-based quantized ML models with YouTube's custom accelerator ASIC (Argos) for video transcoding by converting models into graphical representations for silicon protobufs and running quantization aware training
  • May 2025 - Aug 2025
    Microsoft
    Software Engineering Intern
    • Enhanced one of 9 official AI Foundry customer templates by integrating a multi-agent conversational framework via Semantic Kernel, leveraging custom agents with Azure AI Translation and Azure Language Services (CLU, CQA)
    • Open-sourced contributions on GitHub (60+ stars, 50+ forks) with 400+ customer provisions on Azure in a month
    • Collaborated with the Semantic Kernel team to identify and fix bugs in agent orchestration techniques
  • May 2024 - Present
    Autodesk
    Machine Learning Engineering Intern
    • Developed a Wiki security screening pipeline using GPT-4 that processed over 15,000 pages to identify confidential information such as API keys, usernames/passwords, and IP hostnames
    • Finetuned and deployed text classification models (DeBERTa and Llama3) using AWS Lambda to automatically route Jira tickets, reducing engineer overhead by 80% per ticket
    • Selected among all Autodesk interns to demo both projects in a fireside chat with the Chief Information Officer
  • June 2023 - Aug 2023
    Cubic Transportation Systems
    Software Engineering Intern
    • Developed a full-stack web application for the Boston MBTA transportation system using Flask and React to monitor over 300 devices, as well as various sales and ridership statistics

Publications

  • 2024
    • C. Kymn*, S. Mazelet*, A. Ng, D. Kleyko, and B. Olshausen. Compositional Factorization of Visual Scenes with Convolutional Sparse Coding and Resonator Networks. In Proc. of Neuro Inspired Computational Elements (NICE) Conference

Honors and Awards

  • 2023
    • Cal Alumni Association Leadership Scholar (2023, 2022)
  • 2023
    • Google Cloud Next Student Innovator

Projects

  • N3CTAR - Neural 3D Cellular Tessellated Automata Rendering
    • Developed a 3D triangle mesh to 3D voxel pipeline utilizing Barycentric coordinates and FloodFill
    • Implemented a real-time interactive renderer for a trained 3D convolutional neural network (simulating neural cellular automata) to enable user-driven voxel destruction and regeneration
  • Provable Robustness for Deep Classifiers
    • Implemented the Fast Gradient Signed Method (FGSM) adversarial approximation attack on a trained MNIST classifier and created a robust training regime by optimizing over the non-convex dual value
  • Generating New Yorker Cartoons with Stable Diffusion
    • Finetuned stable diffusion using Dreambooth to output images in the style of New Yorker cartoons