Research
Current — Aesona
Open source
- torchtune Co-founded PyTorch's post-training library; helped scale it to support Llama 2 through 4.2024pytorchpost-trainingllm
- torchmultimodal Co-founded PyTorch's library for training multimodal models at scale.2022pytorchmultimodal
- The Llama 3 Herd of Models Co-authored the technical report for Meta's Llama 3 model family.2024llmllama
Past research
- Thalamocortical Connectivity and Autism Symptoms First-author study of resting-state brain connectivity in autism, done as an NSF Graduate Research Fellow at Stanford.2021
- Inpainting Cropped Diffusion MRI Using Deep Generative Models First-author work with SRI International using generative vision models to reconstruct missing regions in neuroimaging scans.2020
- Evaluation of Transradial Body-Powered Prostheses Using a Robotic Simulator First-author undergraduate research at UT Dallas — early work applying neural networks to robotic prosthetics.2015