Long Story Short
Personal ProjectA web app that turns any book into a short-form, podcast-style audio summary—written by AI and narrated in seconds.
Jessica Li
I'm a full-stack software engineer with a background in product design and statistics. I care about good UX, thoughtful interfaces, and measuring whether what we build actually works.
I graduated from Harvard in May 2025 with a degree in Statistics and Computer Science, where my thinking was shaped by courses in AI and reinforcement learning, usable interactive systems design, statistical inference, and visualization. I'm especially interested in building AI products that are robust and useful in real contexts. My work sits at the intersection of software, product, and data, including a custom API log explorer for Hume AI and user-facing data export and notification features at Sigma Computing.
Outside of work, I'm interested in probabilistic decision-making (poker), working with physical systems (pottery), and observing real-world behavior (photography).
Projects
Make My Menu
CS 178: Engineering Usable Interactive SystemsAn AI-powered recipe generator that creates customizable recipes from ingredients in your virtual pantry—no grocery run required.
Polar Bears & Climate Change
CS 171: VisualizationAn interactive data visualization exploring how climate change affects polar bear populations—featuring animated migration paths, sea ice trends, and diet analysis.
Research
Explorations in things I find interesting.
Smart Ranger: A MCTS-inspired Approach for Optimal Strategy Approximation
CS/STAT 184: Reinforcement LearningDeveloped a modified Monte Carlo Tree Search algorithm with domain-specific pruning heuristics for a sequential card game, improving win rates from 39% to 51% over greedy baselines.
Claim Denials: The Health Insurance Market from a Differential Learning Perspective
ECON 1071 / CS 37: Incentives in the WildExtended classical adverse selection models with dynamic Bayesian learning to analyze how claim denial rates influence consumer dropout behavior and insurer market equilibria.
Breadth of Life: Trees as Bioindicators of Air Quality
STAT 139: Linear ModelsApplied OLS, LASSO, Ridge, and tree-based regression methods to evaluate urban tree characteristics as predictive bioindicators of air quality across 46 U.S. cities.