Stellonlabs is an AI research lab that creates ultra‑compact neural network models designed to run on edge devices such as smartphones, wearables, and IoT sensors. Their micro‑models, often under a megabyte, use quantization and pruning to retain high accuracy while minimizing latency, memory, and power consumption, and are delivered as ready‑to‑deploy packages or APIs compatible with ARM, RISC‑V, and microcontroller platforms.
Funding
$500K 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.
Founders
Product
Problem
Deploying state-of-the-art AI models on edge devices such as smartphones, wearables, and IoT sensors is hindered by high computational, memory, and power requirements, limiting real‑time inference and offline functionality.
Solution
Stellonlabs focuses on creating highly efficient, compact neural network models that retain strong performance while fitting within the strict resource constraints of edge hardware. Their research produces model architectures and optimization pipelines that reduce parameter count and inference latency, enabling on‑device AI without reliance on cloud connectivity. The lab provides these tiny models as ready‑to‑deploy packages or APIs that can be integrated into existing edge software stacks, allowing developers to add sophisticated AI capabilities while preserving battery life and meeting latency targets.
Target Audience
Primary customers are edge device manufacturers, IoT solution providers, and application developers who need high‑performance AI that can run locally on constrained hardware.
Features
- Custom-designed micro‑models with sub‑megabyte footprints for vision, audio, and sensor data processing
- Quantization and pruning techniques that maintain accuracy despite aggressive size reduction
- Hardware‑agnostic deployment libraries supporting ARM, RISC‑V, and microcontroller platforms
- Low‑power inference pipelines optimized for real‑time operation on battery‑constrained devices
- Open‑source reference implementations and documentation for rapid integration