Wayne Hendricks

About

Fri, 08 May 2026 00:00:00 +0000

I keep large training runs alive.

For the last seven years that has meant pathology foundation models at Paige.AI and now Tempus. Before that, it meant the Compact Muon Solenoid experiment at CERN’s Large Hadron Collider, where I ran the Caltech Tier-2 — one of the U.S. compute sites the physics collaboration depends on to reconstruct collision events. Before that, enterprise systems at Duke and Unix production at FedEx.

The thread is not “AI” or “physics.” It is production infrastructure for missions where the cost of a failed run is high — months of physicist time, or weeks of GPU-hours on a billion-parameter model. The skills transfer better than people assume. Scheduling, storage tiering, network fabric, observability, change management under pressure — the LHC people and the foundation-model people are solving the same problem at different scales of data and different shapes of compute.

CV

Fri, 08 May 2026 00:00:00 +0000

Thomas Wayne Hendricks
#

GitHub · Google Scholar · ORCID · twh@waynehendricks.com

Professional Summary
#

Principal HPC engineer specializing in foundation model training infrastructure for clinical and scientific AI. Currently a Senior HPC Engineer at Tempus in New York City, where I lead GPU and storage infrastructure behind histopathology foundation models. Joined Tempus through the September 2025 $81M acquisition of Paige.AI, where I had been the sole HPC engineer from 2018 through the acquisition — scaling the company from 5 to 40+ ML users and helping fill the early engineering staff. Previously principal administrator of the Caltech CMS Tier-2 cluster for the Large Hadron Collider, operating 7,300 HTCondor slots and 4.5 PB of storage in production collaboration with CERN, Fermilab, and the Open Science Grid.

Publications and Acknowledgements

Fri, 08 May 2026 00:00:00 +0000

Papers I co-authored and papers where I built the underlying production infrastructure. The two are intentionally separated — author lines and acknowledgements credit different kinds of work.

Acknowledgements
#

Papers where I am credited for infrastructure contributions rather than listed as an author.

Virchow2: Scaling Self-Supervised Mixed Magnification Models in Pathology
#

E. Zimmermann, S. Liu, et al. · arXiv preprint, 2024

Hello World

Tue, 07 May 2024 00:00:00 +0000

Starting to write posts about topics I like.

Wayne Hendricks

About

CV

Thomas Wayne Hendricks #

Professional Summary #

Publications and Acknowledgements

Acknowledgements #

Virchow2: Scaling Self-Supervised Mixed Magnification Models in Pathology #

Hello World

Thomas Wayne Hendricks
#

Professional Summary
#

Acknowledgements
#

Virchow2: Scaling Self-Supervised Mixed Magnification Models in Pathology
#