In 2018 I joined Roam Analytics as an NLP engineer, where I have been working on improving existing NLP pipelines and developing new models for information extraction applied to clinical text.
I am particularly interested in Natural Language Processing, and over the years gained some experience in python and TensorFlow (see for instance my seq2seq implementation).
I did a MSc in Computational and Mathematical Engineering at Stanford University. I also have a MS in Applied Mathematics from the Ecole polytechnique (France) where I was also involved in statistical and quantum physics.
I did some research with the Machine Learning Group on Natural Language Processing, and was one of the authors of a paper introducing a simple noising technique for data generation applied to the Grammar-Correction task. I invite you to test a cool alternative to grammarly, named Crio, being developed on top of this work.
I also did a project with the SLAC Machine Learning Group during Winter 17 on applying computer vision techniques for 3d cluster-splitting in ATLAS.
During my time at Stanford, I was also a TA for some graduate level classes in Machine Learning :
- CS224n: Natural Language Processing with Deep Learning taught by Christopher Manning and Richard Socher for which I contributed to the lecture notes.
- CS234: Reinforcement Learning taught by Emma Brunskill, for which I developed the assignment on Deep Reinforcement Learning, replicating DeepMind’s DQN paper on Pong.
- CS229: Machine Learning taught by Andrew Ng and Dan Boneh, for which I wrote the lecture note on Linear Quadratic Regulation.
- CS230: Deep Learning, taught by Andrew Ng and Kian Katanforoosh, that follows deeplearning.ai. With Olivier Moindrot, we developed a starter code in Tensorflow for NLP and Computer Vision to help students bootstratp their projects.