Redefining the limits of Artificial Intelligence

Every major breakthrough in AI started as a vision. Our research could be the next big leap.

IA Simbólica

ARQS. Road to synthesis inteligencies (SIs)

Current LLMs (Large Language Models) with billions of parameters demonstrate their immense potential day by day with new evolutions, optimizations, and extremely effective adjustments in specific domains. Nevertheless, at Amgor, we believe that there are different technological spaces yet to be explored, whose potential in areas such as knowledge understanding and dynamic learning adaptation could be several orders of magnitude superior to what is currently happening with LLMs. We are working on developing a new paradigm: Synthesis Intelligences. Our ARQS (Symbolic Architecture for Conscious Emulation) aims to introduce a complementary vision: systems that not only generate responses based on LLM models, but also learn, understand, and continuously optimize their decision-making process in productive environments. With a focus on dynamic Machine Learning and a processing speed that until now can only be achieved with enormous hardware investments. The objective? To have systems whose alignment is to improve each day they are in production, guaranteeing a constantly evolving adaptation and precision.

Much more than sensors. A new approach to distributed intelligence.

As a second line of research, we are working on developing a new way to understand the real-world deployment of AI systems, with a special focus on Edge AI or peripheral intelligence. Most centralized AI systems rely on a single processing source, which limits their ability to respond and adapt in dynamic environments. Our MIDAS approach (Intelligent Distributed Mapping on Symbolic Architecture) is inspired by the architecture of the human brain: a distributed system where each perception unit not only collects data but also processes it and generates local decisions before interacting with the AI central core. By integrating NVIDIA Jetson devices, among others, we enable these nodes, positioned as "active sensors", to function like different brain modules—vision, memory, coordination—allowing for a faster, more efficient, and adaptive real-time response. We are fully aware of the scale of this challenge, but advancements in RAG techniques and AI agent flows have opened up a new frontier. Our company is working to push this to the next level—where efficiency and cost are perfectly aligned.

Percepción Inteligente

WAI. Weareable Artificial Intelligence

WAI is the way we call to those AI services that would be used over an large market of wearable gadgets in a near future. We are currently witnessing the full emergence of IoT (Internet of Things) technologies, enabling the development of miniaturized, interconnected devices applicable to a vast range of processes and services. Our company is focused on developing WAI, which represents a step forward in the creation of interconnected devices—endowing them with intelligence and autonomy (with or without an Internet connection). This advancement is not limited to everyday devices but also extends to new miniaturized devices aimed at the healthcare, industrial, and entertainment sectors. From sensors capable of interacting with enzymatic microarrays to biometric devices that intelligently interact in real time, WAI pushes the boundaries of precision medicine—offering personalized treatment and monitoring at a level that, until now, was pure science fiction. The greatest current challenge in providing true artificial intelligence capabilities to low-power devices with limited computational capacity lies in the size of AI models and the hardware requirements, which today remain unfeasible for portable devices and gadgets. Our company is not only researching domain reduction and AI model distillation to create solutions with lower hardware requirements, but it is also advancing in the development of a specialized version of ARQS and MIDAS, specifically designed for WAI. This innovation will enable the execution of AI services, both online and offline, on common mobile devices.

WAI

Let’s talk about the future of AI.

If you share our passion for innovation and want to explore potential collaborations in our research areas, we would love to hear from you. From technological synergies to new applications, we believe the future of AI is built together.

Contact us

Asistente Virtual

¡Hola! ¿En qué puedo ayudarte hoy?
Tecnología de asistente virtual desarrollada por Amgor
Thinking
Hold on...