Ziroh Labs’ CPU-first AI platform Kompact AI is designed for high-volume enterprise use cases where workload scale, I/O patterns and smaller AI models make efficient CPU-based inference a practical choice
From Left to Right - Prof. S. Sadagopan, Founder Director, IIIT-Bangalore , Hrishikesh Dewan, Co-founder and CEO of Ziroh Labs ,Dr. Whitfield Diffie, pioneering cryptographer and recipient of the 2015 ACM A.M. Turing Award.
Bengaluru: Ziroh Labs, an AI and deeptech startup, has extended the Kompact AI platform across different sectors and form factors.
Ziroh Labs’ CPU-first AI platform Kompact AI is designed for high-volume enterprise use cases where workload scale, I/O patterns and smaller AI models make efficient CPU-based inference a practical choice.
Alongside large, compute-intensive models, organisations are deploying AI across thousands of operational workflows, from software testing and document processing to customer interactions, enterprise automation, cybersecurity and industry-specific applications. Many of these workloads rely on smaller models, process high volumes of data and transactions, and are shaped as much by I/O, latency and deployment scale as by raw compute requirements. Kompact AI is built for this part of the AI landscape: enabling efficient inference for enterprise workloads that can run effectively on CPU infrastructure.
The platform is being taken into multiple sectors through purpose-built Kompact AI Box solutions, packaged AI infrastructure designed around specific industry requirements. These solutions bring together AI models, inference capabilities to address practical use cases across areas including healthcare, education, retail, fintech and sovereign deployments.
Ziroh Labs is working with academic institutions to establish Centres of Excellence (CoEs) focused on AI, giving students access to practical AI infrastructure and environments where they can experiment, build, and deploy AI applications.
These CoEs are intended to create a pipeline of developers, researchers, and entrepreneurs who can work with real AI infrastructure, explore new use cases and contribute to India's growing AI ecosystem.
Ziroh Labs also partnered with academicians from reputed institutes such as IIT Guwahati to develop the Math Lab initiative. The charter of the Math Lab is first to identify whether the mathematics currently employed in large language models and state-space models constitutes numerical structures that can genuinely lead to general intelligence.
Ziroh Labs has collaborated with IIT Guwahati to develop a comprehensive 40+ hour course designed to democratise advanced AI and Large Language Model education for students, developers, researchers, and professionals worldwide. While the course is free for all learners, they would be required to pay for a certification by IIT-Guwahati.
( Source : Deccan Chronicle )
Next Story