May 2024
Joined Harvard Kirchhausen Lab
ResearchCompleted
ML researcher at Harvard Medical School applying 3D vision transformers to cryo-electron tomography. The lab images subcellular structures at near-atomic resolution; my job was to make the resulting volumes interpretable at scale.
● What I shipped
- Trained on multi-node DGX clusters: A100 / H100, NVLink intra-node, Infiniband inter-node, RAID + NVMe storage tier.
- PyTorch FSDP with bf16 mixed precision and activation checkpointing to fit large 3D ViTs.
- Diagnosed a Rendezvous (RDZV) backend issue affecting Infiniband multi-node training — filed PyTorch issue #144779.
● Stack
PyTorchFSDP
CUDAInfinibandNCCLDGXcryo-ET