About
Hi, I'm Linda! I study computer science and classics at Stanford.
I love thinking about how people think and how machines think. I also love finding connections between people, ideas, and things, and I want to build systems that do this kind of knowledge organization at scale.
Recent work:
Selleb — Implemented CLIP embedding pipeline using HuggingFace Transformers + FastAPI, deployed on AWS EC2; built first production recommender system (top-K cosine similarity on CLIP embeddings) for personalized product and user recommendations; explored GNN-based link prediction architectures; deployed backend with ECS + nginx for reliable scaling.
Stanford AI Lab — Trained language models on synthetic data to improve implicit reasoning; implemented GPT-Neo pretraining pipeline with PyTorch + HuggingFace; generated chain-of-thought datasets.
Stanford School of Engineering — TA for CS106A/B (intro Python & C++), leading weekly sections, debugging, grading, and teaching core CS concepts (pointers, objects, recursion, algorithmic analysis, data structures, graphs).
Research interests: RLAIF/RLHF methods for complex reasoning, multimodal retrieval & recommendation, and post-training behavior shaping through synthetic data and in-context learning.
In my free time, you can probably find me reading, jogging, or going to a hackathon.