For developers, ogre.run provides a platform to automate dependency management for source code and deploy the fully functional runtime environment. Unlike traditional devops tools that rely on humans to do the heavy lifting of curating a list of dependencies, ogre.run has AI at its core, saving hours of users’ time when trying to run source code.

Permion.AI produces neurosymbolic AI and security solutions for the hardest problems, spanning banking, enterprise risk, defense. Our platform enables applications in artificial intelligence/machine learning, yielding breakthrough speed, quality and affordability.

Geometric Computing for Artificial Intelligence.Ogarantia is a consulting company that builds software and solutions to improve performance of AI systems. Our goal is to make the process of deploying AI simpler, allowing for engineering teams to continuously achieve their performance goals with less time and effort.Ogarantia ceased operations in France in March 2024 in favor of its US spin-off Ogre.run, Inc.

Upstride enables developers to optimize the performance of neural networks (NNs) via the use of Geometric (Clifford) Algebras. Upstride's product is a computational engine that leverages Geometric Algebra (GA) and serves as backend for Tensorflow, enabling developers to design, train, and deploy NNs that are more compact, i.e., they have a lower count of free parameters when compared to the ones obtained with standard Tensorflow. The product is fully compatible with Tensorflow 2 and works seamlessly with Nvidia's CUDA and Intel's OneAPI. Moreover, the technology is compatible with other standard NN compression techniques such as quantization and pruning. As a co-founder and CTO, my role is to build the technical vision, improve the tech and product roadmap, and more importantly, strengthen the team culture in order to overcome the great challenges that come with building and selling a deep-tech product. Part of the Nvidia Inception program and the Microsoft AI Factory.upstride.io

Leveraging my Ph.D. and signal-processing expertise to found a deep-tech startup.

Research and development of machine-learning algorithms for optimization of High-Performance Computing (HPC) systems; Statistical analysis of HPC performance logs. The algorithms I design are in the core of UCit's products, e.g., Analyze-IT and Predict-IT.

Design of artificial-intelligence algorithms for data partitioning in high performance computing platforms composed by heterogeneous devices (CPUs and GPUs). Main result: a strategy based on adaptive filters for mapping computational kernels onto CPUs and/or GPUs at run time. Supported by a Marie Curie Research Fellowship of the INFIERI network.

Research on adaptive filtering theory, one of the building blocks of artificial intelligence and machine learning. Design and analysis of new adaptive algorithms and structures, as well as their feasibility for real-world applications such as system modeling and signal enhancement. My thesis uses the theory of Geometric Algebra (GA) to formulate a new class of adaptive filters (AFs), the so-called Geometric-Algebra Adaptive Filters.

Research focused on the theory of adaptive filters (one of the building blocks of artificial intelligence). Those adaptive filters are used in signal processing pipelines to model systems at real time. Thesis title: "Incremental Strategies in Combination of Adaptive Filters".

Year-long research stay at the Institute for Media Technology (LMT) of TU Munich. I developed adaptive filters (one of the building blocks of artificial intelligence), based on Geometric Algebra, for 3D registration of Point Clouds.

Design, simulation and verification of analog and digital blocks for application-specific integrated circuits (ASICs). Use of Cadence tools (ADE, AMS designer, Spectre, SpectreRF)

Specialization in integrated-circuits (IC) design and microelectronics sponsored by the Brazilian government and Cadence Design Systems. I coursed the Analog and Mixed-Signal (AMS) track.

Design and simulation of electroacoustic systems. Signal processing for sound engineering.