Thomas Gauthier Email & Phone Number

New York, Ny, Us

Cupertino, California, Us

Usa, Us

• Founded the Hux Data Science Group - Adapted the Skills Rubric and Career Ladder to Product Development roles - Designed the Interview Loop and closed 9 new hires - Created an 8-week Data Science bootcamp enabling Individual Contributors from different backgrounds to share a common set of knowledge• Re-architected Machine Learning stack - Researched Feature Stores (Tecton / Feast), Containerized Workflows (MLFlow / Kubeflow), AutoML (Data Robot / H2O) and Model Deployment (Algorithmia / KF Serving) solutions - Prototyped most promising architectures and studied scaling properties of the different stack configurations to de-risk technology choices• Tech Lead for 2 Data Science teams - Optimization: Leverage Online ML (Reinforcement Learning, Bandits) to tackle the cold start problem - Insights & Analytics: Leverage Offline ML to transform marketing cashflows into increases in Customer Lifetime Value

New York, Ny, Us

• Managed a team of 4 Data Scientist• Engineered an event-based system (HBase) to synch the output of modeling pipelines with The Trade Desk, opening a new channel to monetize our AI and generating 500k in revenue• Coached AdOps and Sales to capitalize on Magnetic’s AI capabilities and differentiate in the marketplace• Spearheaded the Data Science transition activities resulting from Deloitte Digital’s acquisition of Magnetic

New York, Ny, Us

• Code-reviewed user propensity modeling pipeline (XGBoost) that valued 350M live profiles daily• Upskilled on Platform Engineering to implement Machine Learning features directly into Magnetic’s Real Time Bidding code base and deploy them on our production stack (AWS, Pypy, Micro-services, Distributed systems, Ansible)• Developed Magnetic’s Performance Evaluation framework (A/B testing) to measure the uplift of optimizations across 200 concurrent campaigns per day• Architected and implemented real time pacing (Reservoir sampling) and pricing systems (Control theory) that enabled the doubling of Magnetic’s auction processing capacity from 300k QPS to 600k QPS

New York, Ny, Us

• Joined at the closing of Magnetic’s Series C• Upskilled on Data Engineering to implement Machine Learning features directly into Magnetic’s Big Data code base and deploy them on our production stack (AWS, Python, PySpark, HDFS, Parquet, Luigi, Fab)• Upgraded hourly fraud detection pipelines (Outlier detection) using Magnetic’s Auction, Bid, Impression, Click and Tracking events to reduce post click bounce rate by 20%• Code-reviewed hourly click modeling (Vowpal Wabbit) and calibration pipelines

Norwalk, Ct, Us

• One of the first hires in the Model Risk Management Group• Evaluated Econometric Models deployed to measure Loss Risk across GE Capital’s 500B book of Business• Evangelized a R+SQL stack to replace legacy SAS tooling

Atlanta, Georgia , Us

• Accelerated aircraft design process by using designs of experiments to sample distributed CFD runs and build surrogate models• Implemented Genetic Algorithms & Simulated Annealing to solve non-convex systems design problems• Used Monte Carlo methods to simulate the propagation of uncertainty in aircraft sizing• Self-taught Python, C++ (Open MPI & CUDA) & Pig

Blagnac Cedex, Fr

• Supported the development of the yearly R&D roadmap

Phd Abd, Aerospace, Aeronautical And Astronautical Engineering

Master Of Science - Ms, Aerospace, Aeronautical And Astronautical Engineering

Master Of Science - Ms, Aerospace, Aeronautical And Astronautical Engineering