01 // APPLIED AI
Applied AI
Execution of compact neural networks directly on mobile chips, NPUs, and edge processors. Systems focused on ultra-fast inference, low power consumption, and zero dependency on cloud processing.
Building computational intelligence for the edge.
We research and implement systems at the convergence of Applied AI, Computer Vision, Edge Computing, and Data Sovereignty.
02 // COMPUTER VISION
Computer Vision
Research into visual perception architectures inspired by human neural processing, integrated with proprietary optical engines for real-time capture and decoding of industrial standards and GS1 symbologies (2D Barcodes and GS1 Digital Link).
03 // EDGE COMPUTING
Edge Computing
Autonomous computing happens at the point of capture. We build infrastructure for local assistants, on-device inference, and full offline tolerance, ensuring operational continuity without excessive data transfer costs.
04 // DATA SOVEREIGNTY
Data Sovereignty
Information unification under the sovereign data paradigm, where custody strictly remains at the origin. Canonical architectural models aligned with governmental standards, strict privacy compliance, and corporate inviolability.
Publications & Systems in Operation
ORCID Index ↗Compari: Visual Perception Pipeline and Local Retail Auditing
Production mobile application focused on optical capture, offline-first catalog structuring, and real-time data reconciliation directly at the physical point of contact.
On-Device Optical Decoding in High-Density Environments (GS1 2D)
Investigation into lightweight vision and two-dimensional tracking algorithms operating with reduced latency on consumer hardware.
Quantization and Local Execution of Neural Networks for the Edge
Study focused on neural model compression and direct inference on mobile NPUs to mitigate latency and energy consumption.
Canonical Custody Models and Decentralization at the Point of Contact
Conceptual architecture for sovereign handling of corporate and governmental data, ensuring canonical compliance at the origin.
Technical Standards & Regulatory Framework
Architectural alignment with major industrial transitions, artificial intelligence frameworks, and global data sovereignty mandates.
GS1 Transition & 2D Standards
Progressive global replacement of traditional linear barcodes (EAN/UPC) with two-dimensional symbols (structured QR Code with GS1 Digital Link and DataMatrix) by 2027.
Optical decoding of dense matrices at high throughput requires local perception at the edge. Cloud models introduce unviable latency; our engines are engineered for vision inference directly on the capture hardware.
LGPD & GDPR Governance
Privacy legal frameworks in Brazil and the European Union grounded in explicit consent, data subject transparency, and the principle of strict necessity (data minimization).
Convergence solved by architecture: we prioritize local-first systems where sovereign data remains on the origin machine, eliminating in-transit leakage surfaces and avoiding unnecessary cloud transfers.
EU AI Act Regulatory Framework
The world’s first comprehensive legal framework for Artificial Intelligence, establishing strict obligations for transparency, risk management, and model oversight.
Engineering focused on lightweight, purpose-built neural models for edge computing. Isolated on-device execution ensures deterministic traceability, mitigates operational exposure risks, and meets the most rigorous technical audit criteria.