About

Hi!

I'm Samuel Jenkins, a software engineer with experience across ML Ops and ML engineering. I'm passionate about building machine learning systems to be reliable, maintainable and scalable. I currently work at Ultraleap, a computer vision company deploying hand tracking deep learning models on VR edge devices, where I am responsible for the MLOps system and company-wide infrastructure.

I'm of the opionion that the current state of the ML industry generally underestimates the importance of good tooling when it comes to training and deploying models. The efficiency gains that can be seen from slick orchestration, CI/CD, optimised model training and deployment can save a lot of time and money in the long run, whether that's human time or GPU hours. I'm writing this blog to hopefully share some of the things I've learned in the hope that others can benefit when building their own ML systems, and optimising their end to end workflows.