This company develops specialized Large Language Models (LLMs) and security solutions tailored for O-RAN and 3GPP telecommunication standards. Their ORANSight suite provides automated code generation, specification assistance, and network analytics for O-RAN environments. Additionally, WiLM-Sec offers comprehensive, multi-layered security against prompt injection attacks targeting wireless LLMs.
Funding
Funding not disclosed
Founders
Product
Problem
Existing Large Language Models (LLMs) often lack the specific knowledge and security measures required for Open Radio Access Networks (O-RAN) and 5G networks. This can lead to inefficient network management, code generation errors, and vulnerabilities to prompt-based attacks.
Solution
WiSights Lab develops AI and LLM-based solutions tailored for O-RAN and advanced telecom networks, enhancing network efficiency and security. Their ORANSight suite of LLMs serves as specification assistants, automated code generators, and comprehensive guides for O-RAN and 3GPP standards. To protect wireless LLMs from security threats, WiLM-Sec offers advanced prompt injection detection, real-time threat analysis, and seamless integration with existing wireless infrastructures. WiSights Lab also provides GenAI and telecom consulting services, along with custom R&D solutions for cutting-edge telecom technologies.
Target Audience
The primary target audience includes telecom operators, network equipment vendors, and enterprises deploying O-RAN and 5G networks, as well as organizations seeking to secure their wireless LLM environments.
Features
- ORANSight 2.0: A suite of open-source LLMs (1B to 70B parameters) for O-RAN specification assistance, code generation, and guidance, supporting context windows up to 128k tokens.
- WiLM-Sec: A wireless language model security solution featuring multi-stage prompt injection detection using heuristic checks, advanced language model analysis, and a vector database.
- Automated leak detection and reporting system for identifying potential data leaks, including detailed reports on injection attempts and system responses.
- Seamless integration with frameworks like LangChain, Hugging Face Transformers, and OpenAI’s API for quick and secure deployment.
- Continuous security updates and adaptation to emerging vulnerabilities through regular security patches and enhancements.
- GenAI Consulting: Specialized AI services, including LLM development and integration, tailored for telecom and enterprise use cases.