MetalliSense
From reactive analysis to Real Time alloy control
Created on 28th December 2025
•
MetalliSense
From reactive analysis to Real Time alloy control
The problem MetalliSense solves
In the metal industry, critical losses occur during alloy formation, but existing systems focus only on post-production analysis. Process fluctuations and manual decision-making lead to energy waste, material loss, and high emissions. The lack of real-time optimization increases production costs and slows the transition toward sustainable metal manufacturing.
Challenges we ran into
We solved the OPC client web socket connection issues for continuous data streaming from real time hardware machine - Spectrometer.
We handled with different sort of AI services with different endpoints, and accessed efficiently in required frontend. We used AI orchestration layer to access different models at different endpoints and also a separate endpoint for models to work together as agent to communicate to each others decision.
We stored unlabelled data in the mongodb client to deal with huge dataset of 200000 synthetic data with 30 percent of deviation samples.
Tracks Applied (5)
Creative Use of Kiro
AWS
Requestly
Requestly
ELeven Labs
Eleven Labs
Gemini API
Gemini
Best Use of Auth0
AuthO