I'm a researcher in Palo Alto focused on machine learning and optimization algorithms.
My goal is to apply these ideas towards the study of the natural world, positive advancement of technology, and conservation of natural diversity.
Some things I've worked on include:
- Reinforcement learning (RL) and metaheuristics for the bin packing problem
- Metaheuristics for arithmetic circuit optimization
- Physics-informed RL for discovering unknown biomolecular structures
- Test-time scaling of diffusion models with metaheuristics
- Scalable energy-based language modelling
- Stochastic neural network weights for sparsity and low-bit quantization
- Differentiable sparsity for efficient transformer models
Right now I'm building:
- A new type of generative model that can be efficiently trained to perform chain-of-thought reasoning using continuous vectors instead of discrete language.
- An architecture for infinite-context sequence modelling and continual learning.
If you have questions about my research or want to work with me, reach out. My contact info can be found here.