Christopher Bonnett
Machine Learning Scientist at Intercom
London, England, United Kingdom
Summary
👤
Senior🎓
Top SchoolChristopher Bonnett is a Machine Learning Scientist with 12+ years blending academic cosmology and production ML, currently building ML at Intercom after recent engineering roles at ZOE and Flo Health. He led photometric redshift efforts for the Dark Energy Survey, contributing to its first measurement of cosmic acceleration and managing a key science group of ~20 researchers. He has a strong track record applying Bayesian deep learning and neural nets to large-scale, messy data—producing distance estimates for 165 million galaxies and winning the Euclid data challenge. Equally at home in research and product teams, he’s improved tooling and documentation in open-source projects (notably enhancing an MDN workflow in the Edward probabilistic programming library). Christopher holds a PhD in Cosmology and brings a rare mix of domain science, probabilistic modeling, and hands-on engineering for deploying ML in real-world systems.
12 years of coding experience
12 years of employment as a software developer
Msc, Astronomy, Msc, Astronomy at Leiden UniversityDoctor of Philosophy (Ph.D.), Astronomy and Astrophysics, Doctor of Philosophy (Ph.D.), Astronomy and Astrophysics at Pierre and Marie Curie University
English, French, Dutch, Spanish
Github Skills (13)
neural-network10
machine-learning10
deeplearning-ai10
probabilistic-programming10
deep-learning10
tensorflow10
artificial-neural-networks10
data-science10
python9
bayesian-methods9
keras9
jupyter-notebook8
statistics8
Programming languages (4)
Github contributions (5)
blei-lab/edward
May 2016 - May 2016
A probabilistic programming language in TensorFlow. Deep generative models, variational inference.
Role in this project:
Data Scientist
Data ScientistContributions:5 commits, 4 PRs, 14 comments in 2 days
Contributions summary:Christopher's contributions focused on enhancing a Mixture Density Network (MDN) implementation within the Edward library. They refactored the training loop to separate prediction from training, improved code clarity by adding an empty line, and added a comprehensive notebook tutorial demonstrating MDN implementation with Edward, Keras, and TensorFlow. These changes indicate a focus on model optimization, usability, and educational support for MDN development.
inferenceneural-networksmachine-learningprobabilistic-programmingdeep-generative-models
cbonnett/SkyNet_wrapper
Aug 2014 - Oct 2014
Contributions:115 commits in 2 months
pythonskynetdeep-learningpython-wrappermachine-learning
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.