PRM.DEV designs and implements artificial intelligence solutions for embedded devices and real-time systems. We work with neural networks, computer vision and hardware accelerators, select the optimal model and adapt it to a specific platform. We assess the task, propose implementation options, and account for constraints related to heat, power, memory and the cost of the final product.
What we do
Porting and running a neural network on a specific accelerator (NPU).
Annotating the video stream and delivering ready-to-use data to the user.
Selecting the optimal model, training and fine-tuning.
Running multiple neural networks in parallel on a single device.
Predictive analytics, process control, robots and manipulators, quality control and optimization.
Security & 3D Mapping
Video analytics, access control, event detection, 3D spatial mapping based on cameras and lidars for navigation and monitoring.
Video & Interactive Services
Voice control, content-based recommendations, and streaming video processing for TV and online platforms.
Agriculture
Computer vision for precision fertilizer application, plant health assessment, and analysis of agricultural machinery operations.
Construction
3D maps of construction sites, PPE compliance monitoring, predictive equipment diagnostics, asset movement analysis and loss prevention.
Sports Analytics
From automated AI cameras to training analysis, improved safety and fan engagement.
Technology Stack
Hardware Platforms and SoCs
Renesas DRP-AI, Renesas RZ/V2M
Ambarella CVflow
NVIDIA Jetson, NVIDIA TAO
Intel Myriad VPU, Intel OpenVINO
Google Coral
Lattice sensAI
Frameworks and Tools
TensorFlow
PyTorch
Broadmann17
Visage SDK
AWS IoT Core
Ready to integrate AI into your product? Contact us — the PRM.DEV team will select the architecture, platform, neural network and suitable technology stack to meet your requirements.