TLDR;

I am an AI entrepreneur, inventor, and research engineer based in the USA (Tulsa, OK) since Feb 2023, but with a track record of success in Europe (France, Germany - from 2013 to 2023) and Brazil (born in 1985).

• Check out my newest project: Ogre, a metacompiler that analyzes source code to create the best runtime environment for it.

• I have a small AI consulting business to work with select companies: Ogarantia

• Check out my previous startup upstride. Founded in 2018, raised ~EUR 3M and in 2021 was acquired by Content Square (Paris, France).

• My Ph.D. Thesis: Geometric-Algebra Adaptive Filters

• I like Clifford Geometric Algebras and their potential for improving AI

  • Video ICACGA 2022
  • White Paper
  • openga.org (archive)

CV (short version)

CV (full - R&D details)

Short Bio

I currently serve as Chief Architect for Machine Learning at Permion.ai (contractor), a US-based company building neurosymbolic artificial intelligence (AI) products to leverage information buried deep in enterprise data. I am also the founder of Ogarantia, an European deep-tech consultancy company bridging the gap between groundbreaking mathematical techniques and algorithm design to enable the deployment of robust AI in production.

From 2018 to 2021 I was co-founder and CTO of Upstride, a startup that optimized the performance of neural networks (NNs) via the use of Clifford (Geometric) Algebras. As CTO, I was in charge of building the technical vision, improving the tech and product roadmap, and lead the tech/product teams. Upstride raised over EUR 3 million, of which EUR 1.6 million from venture capital. In July 2021, Upstride was acquired by ContentSquare.

Before founding Upstride, I was an applied scientist (machine learning and signal processing) at UCit, where I was responsible for building their statistical-learning algorithms and software for high-performance computing. In 2016 I earned a Ph.D. in Electronics Engineering (machine-learning and signal processing) from the University of Sao Paulo, Brazil, with a one year-long research period at the Technical University Munich (TU Munich), Germany. From 2016 to 2017, I was a Postdoc at Thales Research and Technology in Paris, France, where he developed AI algorithms to increase the power efficiency of supercomputers.

My interests include AI (connectionist and logic-based), signal processing, Clifford (Geometric) Algebras, adaptive filtering, computer vision, applied mathematics, integrated-circuits design, and high-performance computing.

Reach out: wilder@ogarantia.com

Select repositories

• The Upstride Geometric (Clifford) Algebra Engine for designing neural networks in Tensorflow

• Upstride Engine Tutorials

• Upstride Classification API

• Openga

Select publications

• Lopes, Wilder B.; Lopes, Cassio G. Geometric-Algebra Adaptive Filters. IEEE Transactions on Signal Processing, 2019

• Lopes, Wilder B.; Al-Nuaimi, Anas; Lopes, Cassio G. Geometric-Algebra LMS Adaptive Filter and its Application to Rotation Estimation. IEEE Signal Processing Letters, 2016.

• Al-Nuaimi, Anas; Lopes, Wilder B.; Steinbach, Eckehard; Lopes, Cassio G. 6DOF Point Cloud Alignment using Geometric Algebra-based Adaptive Filtering. IEEE WACV 2016 - Lake Placid, NY, USA.

• Al-Nuaimi, Anas; Lopes, Wilder B.; Zeller, Paul; Garcea, Adrian; Lopes, Cassio G.; Steinbach, Eckehard Analyzing LiDAR Scan Skewing and its Impact on Scan Matching. IEEE IPIN 2016 - Madrid, Spain.

• Lopes, Wilder B.; Lopes, Cassio G. Incremental Combination of RLS and LMS Adaptive Filters in Nonstationary Scenarios. IEEE ICASSP 2013 - Vancouver, Canada.

• Chamon, Luiz F. O.; Lopes, Wilder B.; Lopes, Cassio G. Combination of Adaptive Filters with Coefficients Feedback. IEEE ICASSP 2012 - Kyoto, Japan.

• Lopes, Wilder B.; Lopes, Cassio G. Incremental-Cooperative Strategies in Combination of Adaptive Filters. IEEE ICASSP 2011 - Prague, Czech Republic.

Videos

• Geometric-Algebra Adaptive Filters for 3D Registration:

• CanSat Rocket Launch in Noshiro, Japan: