Eyad Gomaa
CEO & Co-Founder of SILX AI | Building the future of Synthetic Intelligence
Hey there, I'm Eyad Gomaa, CEO & Co-Founder of SILX AI, an 18-year-old AI researcher from Egypt. I'm pioneering the next generation of artificial intelligence through Synthetic Intelligence (SI) systems that transcend traditional transformer and attention-based models. My vision is to build AI that achieves unprecedented data efficiency, real reasoning capabilities, and potentially infinite context windows. I'm the creator of TARS-1B, a revolutionary 1B parameter model based on Liquid Neural Networks that breaks the Chinchilla scaling law, learning with 100×–10,000× less data than transformers. My core innovation, Tau Dynamics (τ Dynamics), enables temporal reasoning to emerge directly in the forward pass without recurrence or attention.
I'm building the future of AI through breakthrough architectures like the Quasar model series with Universal Stage 0–6 Reasoning Pipeline, SI-1 (Synthetic Intelligence-1), and World Language Models (WLMs). My philosophy: current AI are "walking zombies" that simulate thinking but lack real reasoning. My goal is to build a Digital Neocortex that achieves true Synthetic Intelligence.
As an AI researcher and founder of SILX AI, I specialize in cutting-edge artificial intelligence architectures and synthetic intelligence systems. My expertise spans Liquid Neural Networks, continuous-time dynamics, transformer alternatives, and novel reasoning pipelines that push the boundaries of what's possible in AI.
Revolutionary neural architecture with continuous-time dynamics
Continuous-time ODE control parameters for temporal reasoning
AI systems that develop their own kind of thinking beyond human mimicry
Advanced attention mechanisms and infinite context windows
Differential equation-based neural computation frameworks
Structured cognition systems with self-auditing capabilities
Arabic
English
German
My AI research journey represents breakthrough innovations in artificial intelligence. Each project pushes the boundaries of what's possible, from revolutionary neural architectures to synthetic intelligence systems that redefine how machines think and reason.
Founded SILX AI and began pioneering research into Synthetic Intelligence systems that transcend traditional transformer architectures.
Developed TARS-1B, a revolutionary 1B parameter model based on Liquid Neural Networks that breaks the Chinchilla scaling law with 100×–10,000× data efficiency.
Invented Tau Dynamics (τ Dynamics) - continuous-time ODE control parameters enabling temporal reasoning in forward pass without recurrence or attention.
Created the Quasar model series with Universal Stage 0–6 Reasoning Pipeline, introducing structured cognition and self-auditing capabilities.
Pioneered No-Positional-Embedding Transformers research, achieving true infinite context windows through evolving attention mechanisms.
Developed SI-1 (Synthetic Intelligence-1) - experimental architecture implementing the Synthetic Intelligence Principle for real-time adaptive reasoning.
Advanced World Language Models (WLMs) research, extending beyond text generation to interactive environment simulation and embodied intelligence.
Building the Digital Neocortex - a structured, self-organizing reasoning system aimed at achieving true Synthetic Intelligence beyond human limitations.