|
Jerry Han I'm a rising sophomore at Princeton University majoring in Mathematics and Computer Science. I am currently exploring a generative AI startup and conducting research in computer vision at the Princeton Vision & Learning Lab. I am an IPhO gold medalist and veteran of the Singapore Army, where I also received the Soldier of the Year award. At Princeton, I am a member of the Princeton Undergraduate Capital Partners, Association for Computing Machinery and the treasurer of the Malaysian and Singapore Association. In my free time, I enjoy billiards, boxing, card games, foosball and travelling. I avidly follow MMA and I support Real Madrid. |
|
Professional ExperienceMy professional experiences span across a variety of fields, including software engineering, AI research, quantitative research, venture capital and entrepreneurship. |
|
Co-founder of Stealth AI startup, Summer 2024.
Co-founded a startup at the intersection of consumer social and generative artificial intelligence, supported by Sebastian Thrun & Hillspire and mentored by Ameer Haj-Ali & David Charron. I worked on end-to-end development of mobile applications, recommendation algorithms, generative AI & agentic AI systems. I onboarded 106 clients and accumulated >13K views across various social media channels. |
|
AI Researcher at the Princeton Vision and Learning Lab, Spring 2024. Image segmentation and pose detection for object simultaneous localization and mapping (SLAM). Working in collaboration with Jia Deng and Siyang Wu. |
|
Diligence Analyst, Princeton Undergraduate Capital Partners, 2023 - Current. Conducted due diligence analysis which informed investment decisions for a diversified conglomerate’s asset management arm. Completed 7 industry deep dives into ESG integration across multiple asset classes including Fixed Income and Public Equities. Sourced and pitched 2 new AI developer tooling startups which informed investment decisions for a partner at Sequoia Capital. |
|
Quantitative Research Intern, Government of Singapore Investment Corporation (GIC), 2023 Developed a new portfolio construction methodology based on Monte Carlo simulations and scenario analyses, which was incorporated into GIC’s systematic strategies ($360 billion AUM). Constructed a stochastic volatility model to simulate scenarios and price derivatives for Fixed Income and Public Equities. Formulated efficient convex and numerical optimization procedures (15x increase in efficiency) for finding optimal portfolio exposures and trade structures which helped a team of ~30 traders refine their trading strategies. Encapsulated these procedures into purpose-built software packages in R and C++. |
Projects |
|
|
SearchDestroy
Algorithm to sweep an area with multiple drones in the most time efficient way; efficient parametrization of search area and ensuring robustness of algorithm under adversarial interference. Robust multi-drone search algorithm: DARP + heuristics to handle drone loss and online path re-computation. Physics simulation and visualization using AirSim; awarded rank 2nd at AGI House Hackathon (Summer 2024). |
|
Moco App for users to cast a “charm” to protect themselves without having to directly interact with their phone. 3D motion tracking app built in Python, React and SwiftUI for contactless gesture-based interaction with smartphones. Awarded Best Overall Hack at HackPrinceton (Fall 2023). |
|
ANTIDOTE: Artificial Neural Network Trojan Detection Using Topological Data Analysis Estimators Developed a novel robust, scalable and explainable approach to detecting Trojaned neural networks using topological data analysis. In collaboration with Huxley Marvit, Rodrigo Porto and Matthew Banaag. |
|
Algorithmic trading of cryptocurrency futures Traded 12 cryptocurrency derivatives with a monthly volume >$1 million and monthly return of 11%. Statistical analysis of market participants’ trading patterns and market microstructure. Algorithmic trading program in Python to execute trades with 20x leverage with trade frequency of under 5 minutes. |
|
Brainhack Champions Developed a computer vision model (Detectron2 and ModaNet) for clothing identification. Kernel methods for obstacle detection and path-finding algorithm in Python for autonomous navigation of an obstacle course. |
Others |
|
Awards
|
|
Courses
|
|
|