Nathan Zhang

Hi! I'm a 17 yr old High Schooler at Saratoga High School in California. I am a prospective computer science major and am interesting in machine learning.

Currently, I am an incoming summer intern at Aizip, developing TinyMLs for a variety of time-series-based applications.

In the past I've worked for Podium's machine learning team. I also worked on Torus, a social media startup that I built the database and assisted in the backend for.

Email  /  Github  /  Linkedin

Research

Machine learning and deep learning models are potential vectors for various attack scenarios. For example, previous research has shown that malware can be hidden in deep learning models. Hiding information in a learning model can be viewed as a form of steganography. In this research, we consider the general question of the steganographic capacity of learning models. Specifically, for a wide range of models, we determine the number of low-order bits of the trained parameters that can be overwritten, without adversely affecting model performance.

Experience

• Designed and implemented a PostgreSQL database from scratch for Torus, a dynamic social media platform, ensuring optimal data structure and integrity for efficient information storage and retrieval.
• Created the CDN, and collaborated closely with the software development team to integrate the database with the application backend
• Assisted in the development of the backend using node.js

• Finetuned generative AI models to improve the quality, realism, and expressiveness of the podbooks
• Assisted in employing natural language processing (NLP) techniques to interpret transcripts effectively, including languages like Chinese.

• Researched the steganographic capacity of selected machine learning models in order to figure out the capacity of machine learning models possesing malware.

Projects

A Wordle based fencing game for guessing a male epee fencer in the top 61 FIE rankings.

• Python Flask app for detecting and predicting osteoporosis in images and lifestyles using CNN and regression models
• 1st place Hackathon Winner

Template found here.