Here are some of the projects I've worked on. Most are written either to solve my own problems or just for the fun of building things. Check out my GitHub for more.
Showing 3 projects
Designing a full-stack, LLM-powered digital twin that learns an individual's writing style from email history and generates emails in their voice using a modular feature-training-inference pipeline.
Built a cattle breed classification system around ResNet-18, reaching 95% accuracy across 8 breeds and improving robustness on imbalanced data through class weighting and staged fine-tuning.
A running collection of research paper implementations where I break down the intuition, mathematics, and code behind important ML and deep learning ideas.
