Benjamin Kreis
Machine Learning Engineer, Special Projects at Apple
Cupertino, California, United States
Summary
🤩
Rockstar🎓
Top SchoolBenjamin Kreis is a Machine Learning Engineer with 11 years of experience blending physics research and applied ML, currently working on Special Projects at Apple in Cupertino. He transitioned from CERN and Fermilab postdoctoral work into industry ML roles, spending four years building healthcare-focused machine learning at Apple before moving to special initiatives. His background includes a Ph.D. in Physics from Cornell, giving him strong foundations in data-driven modeling, simulation, and quantitative problem solving. Benjamin contributes to open-source FPGA-focused ML tooling (hls4ml), where he implemented and refined CNN support and keras-to-HLS conversion, demonstrating expertise in deploying neural networks to hardware. He is skilled at bridging research-grade models and production constraints, especially for resource-constrained or real-time systems. Colleagues describe him as a pragmatic engineer who brings experimental rigor to production ML and a knack for translating complex algorithms into efficient implementations.
11 years of coding experience
9 years of employment as a software developer
Bachelor of Science (B.S.), Physics, Bachelor of Science (B.S.), Physics at Case Western Reserve UniversityDoctor of Philosophy (Ph.D.), Physics, Doctor of Philosophy (Ph.D.), Physics at Cornell University
Github Skills (8)
mask-rcnn10
keras10
faster-rcnn10
machine-learning10
hlsl10
python10
fpga9
vivado8
Programming languages (5)
Github contributions (5)
fastmachinelearning/hls4ml
Oct 2017 - Apr 2019
Machine learning on FPGAs using HLS
Role in this project:
ML Engineer
ML EngineerContributions:1 release, 148 commits, 27 PRs in 1 year 5 months
Contributions summary:Benjamin's contributions primarily focused on implementing and refining convolutional neural networks (CNNs) within the HLS4ML framework. They modified code related to 1D and 2D convolutional layers, adding parameters, functions and incorporating padding and stride options. The user also made changes to the keras-to-hls conversion process, including adjustments to the model parsing and HLS generation, incorporating a zero counter and integrating the generation process to a common place for use with other ML packages. They updated and tested examples using a Keras-trained CNN.
pytorchpythonvivadofpgasonnx
DisplacedHiggs/NJetsEstimate
Nov 2016 - Apr 2018
Contributions:2 releases, 4 PRs, 28 pushes in 1 year 5 months
backgroundestimate
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.