Thomas Delteil
Senior Machine Learning Scientist at Amazon Web Services
Vancouver, British Columbia, United States
Summary
🤩
Rockstar🎓
Top SchoolThomas Delteil is a Senior Machine Learning Scientist with 11 years of experience building production AI at scale, currently focused on Document AI and Intelligent Document Processing for Amazon Textract at AWS. He blends applied research and engineering—having worked on cloud AI at Microsoft and full-stack R&D at Bloomberg—while contributing to notable open-source projects such as Apache MXNet and Gluon-CV, where he improved tutorials, CI stability, and cross-platform compatibility. Comfortable across the stack, he has hands-on expertise in multi-GPU training, model fine-tuning, and deployment best practices, and often bridges gaps between documentation, tooling, and reproducible ML workflows. Trained as an aerospace engineer with an MSc in Advanced Computing (Distinction) from Imperial College London, he brings a disciplined, mathematically grounded approach and an uncommon background that spans aerospace, cloud AI, and industrial-scale ML systems.
11 years of coding experience
4 years of employment as a software developer
CPGE, Mathematics and physics, CPGE, Mathematics and physics at Clemenceau, NantesFrench baccalauréat, Mathematics, First Class Honours, European Distinction, French baccalauréat, Mathematics, First Class Honours, European Distinction at Lycée Saint-JosephMaster of Science (MSc.), Advanced Computing, Graduated with Distinction, Master of Science (MSc.), Advanced Computing, Graduated with Distinction at Imperial College LondonMaster of Science (MSc.), Aerospace Engineering, Ingénieur SUPAERO, Master of Science (MSc.), Aerospace Engineering, Ingénieur SUPAERO at ISAE-SUPAERO
French, Spanish, English, German
Github Skills (18)
develop10
python10
image-classification10
mxnet10
machine-learning10
gluon10
mask-rcnn10
deeplearning-ai10
deep-learning10
object-detection10
computer-vision10
faster-rcnn10
documentation10
testing9
cicd9
Programming languages (12)
Github contributions (5)
apache/mxnet
Mar 2018 - Nov 2019
Lightweight, Portable, Flexible Distributed/Mobile Deep Learning with Dynamic, Mutation-aware Dataflow Dep Scheduler; for Python, R, Julia, Scala, Go, Javascript and more
Role in this project:
Back-end Developer & DevOps Engineer
Back-end Developer & DevOps EngineerContributions:71 commits, 116 PRs, 35 pushes in 1 year 8 months
Contributions summary:Thomas's primary contributions involve fixing typos and updating documentation related to the MXNet library. They addressed a typo in the `gluon/trainer.py` file and updated the documentation for `ndarray` and `sparse` modules. Further contributions included fixing an issue with tutorial tests by adjusting the test scripts, as well as, fixing a flaky tutorial test in the CI. They also updated the documentation regarding RNN layers.
pythonschedulerdataflowmutationdata-science
ilkarman/DeepLearningFrameworks
Mar 2018 - Jun 2018
Demo of running NNs across different frameworks
Role in this project:ML Engineer
ML EngineerContributions:20 commits, 12 PRs, 11 pushes in 2 months
Contributions summary:Thomas primarily contributes to the development of machine learning models within the repository, specifically focusing on implementing and updating various Gluon-based neural network models for image classification. Their work involves adapting existing models, fine-tuning hyperparameters, and optimizing the training process for multi-GPU environments. The user also incorporates data loading, preprocessing, and evaluation steps within the developed notebooks, indicating a comprehensive approach to model building.
mxnetdeep-learningmachine-learningcaffetensorflow
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.