imply+infer
Building adaptive hardware interfaces for edge AI systems. Our mission is to enable seamless peripheral inference, driver synthesis, and cross-architecture device abstraction for the next generation of intelligent edge computing.
DATASHEET
Rev. 1.0
Nov 2025
Jetson Orin Nano Field-Prototyping Kit
Complete edge AI development system with 67 TOPS performance, dual stereo vision cameras, NVMe storage, and pre-configured software stack. Purpose-built for rapid prototyping of autonomous systems and computer vision applications in professional and research environments.
Core Compute Module
Module:NVIDIA Jetson Orin Nano Super
AI Performance:67 TOPS (INT8)
GPU:1024-core NVIDIA Ampere with Tensor Cores
CPU:6-core ARM Cortex-A78AE @ 2.0 GHz
Memory:8GB 128-bit LPDDR5
Storage:256GB NVMe SSD (M.2 2280)
Power:7W to 25W configurable modes
Thermal Solution:Passive heatsink (fan-optional)
Module Size:100 × 79 mm
Vision System
Camera Sensors:Dual Sony IMX219 8MP modules
Resolution:3280 × 2464 pixels (8 megapixels)
Frame Rate:21 fps @ full res, 30 fps @ 1080p
Pixel Size:1.12 µm
Interface:MIPI CSI-2 (2-lane per camera)
Stereo Baseline:60mm camera separation
Mount:Custom 3D-printed rigid bracket
Features:Pre-aligned, plug-and-play stereo vision
Connectivity & I/O
Ethernet:Gigabit Ethernet (10/100/1000 Mbps)
Wi-Fi:802.11ac dual-band (2.4/5 GHz)
Bluetooth:Bluetooth 5.0
USB Ports:4× USB 3.2 Gen 2 (10 Gbps)
Display:DisplayPort 1.2
GPIO:40-pin expansion header
M.2 Slots:Key M (NVMe), Key E (Wi-Fi)
Pre-Installed Software
OS:Ubuntu 22.04 LTS
JetPack SDK:6.2 with BSP and drivers
CUDA:12.2 with cuDNN 8.9
TensorRT:8.6 for inference optimization
PyTorch:2.1 (ARM-optimized build)
OpenCV:4.8 with CUDA, GStreamer support
Containers:Docker 24.0, NVIDIA Container Runtime
ML Tools:llama.cpp, HuggingFace, ONNX Runtime
Included Components
•NVIDIA Jetson Orin Nano Super Developer Kit
•Dual Sony IMX219 8MP camera modules with mounting bracket
•256GB NVMe SSD with pre-configured system image
•7" portable LCD display (800×480)
•Wireless keyboard and mouse (2.4GHz)
•DisplayPort to HDMI adapter
•19V/3.42A power supply
•3D-printed protective case
•Quick start guide and setup documentation
Performance Reference
YOLOv8 (640×640):~40-50 FPS with TensorRT
LLaMA 7B (Q4):~10-15 tokens/sec
ResNet-50:~300+ images/sec (batch)
Boot Time:< 30 seconds to desktop
Camera Pipeline:Dual 1080p @ 30fps simultaneous
Performance varies by model configuration, quantization, and thermal conditions
Target Applications
•Autonomous robotics and navigation systems
•Computer vision prototyping and development
•Edge AI model deployment and testing
•Research and academic projects
•Industrial automation and inspection
Physical Specifications
Dimensions:210 × 160 × 85 mm (assembled with case)
Weight:~850g (complete system)
Operating Temp:0°C to 50°C (ambient)
Ordering Information
Price:$599 USD (complete kit)
Availability:In stock, ships within 3-5 business days
Warranty:1-year limited hardware warranty
Website:implyinfer.com
© 2025 imply+infer. All specifications subject to change without notice. NVIDIA, Jetson, CUDA, and TensorRT are trademarks of NVIDIA Corporation. For technical support: support@implyinfer.com | Documentation: docs.implyinfer.com