Jeremiah Harmsen
Zurich Site Co-lead at Google DeepMind
Zurich, Zurich, Switzerland
Summary
🤩
Rockstar🎓
Top SchoolJeremiah Harmsen is a seasoned research and engineering leader who has spent two decades at Google and DeepMind driving curiosity-led ML research and production-ready infrastructure. As Zurich Site Co-lead and former Principal Engineer, he builds and coordinates distributed teams across Europe and the US working on computer vision, neural radiance fields, weather prediction, and ML systems. He founded TensorFlow Hub and TensorFlow Serving and created Google’s Machine Learning Ninja rotation, blending open-source impact with internal talent development. His GitHub contributions to flagship TensorFlow repos include improving training loops, dataset tooling, and model-serving integration—work that bridges research prototypes to robust production deployments. With a Ph.D. in electrical engineering from Rensselaer and a track record of founding reusable frameworks, he combines deep technical rigor with people-centered leadership. Outside work he’s based in Zurich and plays volleyball, reflecting a collaborative, team-first approach.
13 years of coding experience
11 years of employment as a software developer
UC Berkeley College of EngineeringPh.D. Electrical Engineering, Ph.D. Electrical Engineering at Rensselaer Polytechnic Institute
English
Stackoverflow
1reputation
0reached
2answers
0questions
Github Skills (15)
keras10
machine-learning10
deep-learning10
tensorflow10
serve10
python10
bert9
modeling9
trainings9
neural-network9
nlp9
callback8
cpp7
android6
tensorflow-serving6
Programming languages (3)
Github contributions (5)
tensorflow/serving
Feb 2016 - Jul 2016
A flexible, high-performance serving system for machine learning models
Role in this project:
ML Engineer
ML EngineerContributions:19 commits, 11 PRs, 8 pushes in 4 months
Contributions summary:Jeremiah focused on improving the Inception model within the TensorFlow Serving system. Their work involved migrating class description translation into the TensorFlow graph for better model hermeticity. They also updated the Inception description lookup code to correctly handle the background class and fixed a typo. This indicates a focus on model integration and refinement within the serving environment.
cpppythonservingdeep-learningml
tensorflow/models
Apr 2020 - Jul 2020
Models and examples built with TensorFlow
Role in this project:ML Engineer
ML EngineerContributions:8 commits in 3 months
Contributions summary:Jeremiah focused on enhancing the model training and evaluation processes within the TensorFlow models repository. Their commits included modifications to the training loop, enabling custom callbacks, and adding functionalities for sub-model checkpointing and final evaluation. Furthermore, they contributed to data processing tools by introducing dataset module import capabilities. Their work also involved adding a new network for per-token classification tasks, extending the model's capabilities for named entity recognition.
deep-learningtensorflow
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.