Mofakir AI

our start-up respond on major question how AI can applied in industrial and delivery augmented generative AI services tha can handel with accuracy and trused result to solve The Efficiency

Comment

Inspiration

This work originated from an independent research initiative developed during the creation of a DeepTech startup, SIRAP, focused on applying generative AI to provide intelligent decision-support for industrial production processes

What it does

The goal is to enable AI systems to understand the operational logic of complex industrial domains by training models on technical and process-specific data, allowing them to identify hidden industrial relationships, production phenomena, and correlations between plant data and quality or resource-efficiency metrics

How we built it

Mofakir is a new approach to representing the meaning of natural text in order to create a mathematical knowledge graph. It not only captures the meaning of natural language but also transforms it into an efficient, structured computational form. This structure can use to detect inference paths, enabling reasoning engines to deliver highly accurate results. The goal is to reduce the risk of deviation caused by relying solely on probabilistic models when leveraging the full potential of large language models (LLMs). This allows us to apply LLMs more reliably in exact science fields such as medicine, healthcare, industrial automation, and beyond.

Challenges we ran into

A central challenge is that conventional AI models rely on text-to-text understanding, while industrial data is inherently relational, dynamic, and tightly coupled with automated control programs and underlying physical phenomena. Traditional mathematical and model-based control approaches have repeatedly failed to fully automate industrial processes because they treat the problem strictly as a computational one; any unmodeled external factor such as climate effects, machine aging, or emergent physical behaviors degrades performance significantly

Accomplishments that we're proud of

we have MVP can show the concept and framework call AlgSubGr* is Knowledge-Augmented Language Models.

What we learned

the LLMs in the futur can change everything in automation world begin form control systeme to plc

What's next for Mofakir AI

in the finale phase will have automation platform changing the world of plc and scada by AI agent delivery AI assistance to user with different level (finance,technical,staff,shareholding....) from data control streaming and users question can imagine the real industrial process and drawing all control entity that help AI to assists the human with efficiency.

Built With

Share this project:

Updates