Nicolas Gay
Co-founder @Nabla Applied Research
Signup · Get unlimited contacts
WORK HISTORY
Co-founder @Nabla Applied Research
Vigo, DE
Full-stack Architecture and Implementation: • End-to-end architecture and implementation of a docker-based platform including satellite and sensor data collection, Kalman-based data fusion, digital-twin service, GPU-accelerated simulation, AI based recommendations and full-featured web-based frontend. • React-based client-side web frontends for remote sensing, data fusion visualization, simulation (single and ensemble simulations), dataset generation, and visual DNN training, testing and validation. GPU-accelerated visualization (3D/2D). • AWS / Strato-based dockerized backend solution using Python FastAPI. • DevOps technologies (Docker, Docker Compose, Docker Swarm,Prometheus) and various databases (PostgreSQL, Redis) for scalable deployments and high throughput.Machine Learning: • DNN dataset creation, DNN training and validation for physics-inspired DNN types, e.g. Fourier Neural Operators(FNOs), using PyTorch and GPU (Apple MPS, Nvidia CUDA) support. • Implementation, training and validation of Physics-Inspired Neural Operators (PINO), including architectural optimization and hyper parameter tuning, as well as GPU memory footprint optimization through Tensorized FNOs (TFNOs). • Implementation of C++ surrogate solvers based on FNOs for physics-based (FEM/FVM) simulators. • Implementation of C++ DNN trainer for Apple Metal Performance Shaders (MPS) running natively on MacOS
EDUCATION
Instituto Tecnologico de Buenos Aires (ITBA)
Electrical Engineer, Communications and control systems
Technische Universität Dresden
Master of Science (M.Sc.) / Diplomingenieur (Dipl.-Ing.), Semiconductor technology, microelectronics, integrated circuit design
Technische Universität Dresden
Doctor of Philosophy (PhD), Analog circuit design with organic field-effect transistors
SKILLS
ABOUT NICOLAS GAY
I like challenges of any kind, I thrive in uncertainty, and I am creative and resourceful. Just throw me a bone to bite on and I will deliver.Last year, we founded Nabla Applied Research to solve a fundamental challenge in scientific computing: physics-accurate simulations for real-time applications. Nabla\'s core technology replaces the classical FEM/FVM solver engine found in scientific computing simulators with a Machine Learning-based surrogate model that preserves the underlying physics while running orders of magnitude faster, at a fraction of the cost.We recognized immediate needs for accurate real-time simulations in climate change scenarios: Everywhere around us climate emergencies are more intense and dreadful with every season, producing avoidable damages and human losses. Therefore, we created Qinetiqs, Nabla\'s dedicated platform for climate emergency operations and risk prevention. Qinetiqs provides real-time multi-hazard intelligence for floods and wildfires, serving civil protection agencies, emergency responders, and insurance companies and infrastructure operators across Europe.We are looking for investors, EU grant partners, and pilot partners in civil protection and risk prevention/mitigation. → n••••••••@nabla-research.comhttps://nabla-research.com/qinetiqs
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.