Projects
I am a huge proponent of project-based learning and am always working on something. I started prioritizing projects during graduate school to solidify my knowledge and application of machine learning, and I have found them to be a great way to learn new concepts, sharpen existing skills, and explore curiosities that fall outside of my work areas.
Perfect Form
Applying computer vision models for 3D pose estimation to analyze powerlifting mechanics.
- Perfect Form: Computer Vision for Powerlifting
- Perfect Form: To the Cloud
- SAM 3D
- Scaling Local Compute (Second-half)
- 3D Pose Harness
Local LLMs
Training modest large language models on consumer hardware from scratch.
- Nanochat
- Building Large Language Models
- Building a Tokenizer from Scratch
- Implementing Flash Attention
Lost Cities
Learning optimal strategy in an imperfect-information card game using reinforcement learning.
- Lost Cities: Using Deep Q-Networks to Learn a Simple Card Game
- Lost Cities Part 2: Learning from Deep Q-Networks
Steep
Training neural networks for real-time navigation in video game courses using imitation learning.