LGN provides edge AI management software to orchestrate deployments and supervise models on resource-constrained edge devices. This platform enables continuous learning and optimization of AI systems operating with real-world sensor data. The solution helps organizations scale commercial edge AI products while controlling capital and operational expenditures.
Funding
$4M raised to dateRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
IIFounders
Product
Problem
Scaling edge AI deployments presents challenges in model supervision, data management, and cost control, hindering efficient product deployment and commercial viability. Monitoring model performance in diverse real-world conditions and managing data transfer volumes from edge devices to the cloud can lead to escalating operational expenses.
Solution
LGN provides an edge AI management software platform that enables technical and financial oversight of edge AI deployments. The platform orchestrates deployments using Kubernetes and OpenShift, manages AI model deployment across fleets of edge devices, and supervises model performance with real-world data. It optimizes data selection to reduce transfer and processing costs without sacrificing visibility or learning speed. By extending cloud orchestration to the edge, LGN allows companies to scale edge AI products while maintaining control over costs and ensuring reliable operation in real-world environments.
Target Audience
The primary target audience includes enterprises deploying AI at the edge in industries such as automotive, manufacturing, and IoT, as well as data scientists and machine learning engineers responsible for managing and optimizing edge AI models.
Features
- Orchestration framework extending Kubernetes and OpenShift to the mobile edge
- AI/ML model deployment management across large fleets of edge devices
- Real-world model performance monitoring and improvement through continuous retraining
- Data selection optimization to reduce data transfer, storage, and processing costs
- Supervised edge device runtime with scheduler client, supervisor, and MQTT broker for sensor integration
- Configurable workflows for data selection and Kubeflow re-training
- Real-time fleet monitoring with alerts for performance drops
- Automated version control and rollbacks for deployed models