Somerset, NJ —September 3, 2025 — VCTI, a leader in broadband and telecom network intelligence, has released a first-of-its-kind open-source AI model designed specifically for the optical communications and photonics industries.
The new encoder model developed at VCTI, VCTI-RoBERTa-Fiber, is trained on highly specialized datasets from the fiber optics sector, making it uniquely capable of understanding the industry’s complex technical language. It is now freely available on Hugging Face for developers, researchers, and equipment manufacturers.
“Most Encoder AI models are trained on general internet text — great for everyday language, but not for the precision and depth our industry demands,” said Manuj Nayak, Head of the AI Center of Excellence, VCTI. “VCTI-RoBERTa-Fiber changes that. It’s a purpose-built AI model that can jumpstart product innovation, speed up development cycles, and enable smarter, AI-powered tools for optical networks. By making this model open-source, we’re inviting collaboration across the optical communications ecosystem. We believe shared innovation accelerates progress for the entire industry.”
Designed for Real-World Impact
The release builds on VCTI’s AI initiative, which focuses on creating domain-specific intelligence tools to advance telecom innovation.
The model is available now here.
About VCTI
VCTI offers a suite of SaaS-based applications that unlock the potential of service providers’ broadband expansion investments. The company uses cutting-edge AI-driven data analytics to pinpoint regions with strong growth potential, while uncovering hidden opportunities within the existing footprint. This allows service providers to make business decisions quickly and decisively. A trusted partner to the most respected and innovative broadband and technology companies worldwide, VCTI is a privately held, global company headquartered in the US with offices in India. Learn more at www.vcti.io