I'm a polymath, AI researcher, applied mathematician, and innovation architect.
My work explores the frontiers between symbolic and sub-symbolic reasoning, geometry and cognition, and classical and quantum computation.
I believe in human-augmenting AI – not to replace us, but to deepen our capacity to think, act, and evolve.
- Transformers and Explainability – Spectral pruning, random matrix theory, SVD analysis
- Quantum Information and Geometry – Weak values, superoscillations, Qiskit simulations
- Mathematical Structures in AI – Simplicial sets, Clifford algebras, category theory
- Cognitive AI Architectures – LangChain agents with long-term memory and reasoning chains
- Generative Intelligence – Tools that elaborate raw ideas into structured mathematical form
-EE4C2C?style=for-the-badge&logo=pytorch&logoColor=white){kind=icon}
A modular Python library for category theory, including:
- Categories, functors, and natural transformations
- 2-categories, toposes, monoidal structures
- Simplicial sets, groupoids, presheaves, infinity categories
NeuroGlyph is a new protocol language designed for structured communication among cognitive agents, both human and artificial.
Built around the core ideas of relation, context, and semantic bridges, it's intended for:
- Modeling complex interactions in collaborative environments
- Powering tools like HU-Core and multi-agent systems
- Enabling philosophical dialogue, collaborative writing, and creative research sessions
“Every relation is a cognitive act. NeuroGlyph provides its grammar.”
| Repository | Description | Topics |
|---|---|---|
NeuroScan |
Visualizing neural network activation maps layer by layer | PyTorch, Interpretability, Visualizations |
WalkGPT |
Merging knowledge graphs and LLMs via guided random walks | Knowledge Graphs, LLM, Prompt Engineering |
DeepSpectra |
Applying Random Matrix Theory for deep networks spectra | RMT, Deep Learning, Diagnostics |
CliffordLab |
A lab to experiment with Clifford Regularity | Clifford Slice Regularity, Clifford ML |
Other projects are currently under private development, including a multi-agent system for mental, physical, and emotional transformation through AI-enhanced self-knowledge. If interested contact me directly.
"Mathematics is not a tool for science. It is the language of Reality itself."
I'm interested in how high-level abstractions can guide intelligence, both human and artificial. I explore connections between phenomenology, formal logic, and neural computation, with a pragmatic engineering attitude.
- Multi-agent architectures for generative reasoning
- Quantum-classical hybrid inference models
- Active perception and attention-based learning