ZedIoT engineering blog

AI + IoT Engineering Insights

Practical guidance for teams building connected products: from embedded firmware and edge gateways to IoT platforms, industrial integration, and applied AI.

Technical library

Browse by engineering topic

Go directly to the area that matches your current architecture, device, platform, or product decision.

Latest updates

Recently published articles

New technical comparisons, implementation notes, and field-oriented engineering guidance.

Generated embedded product development scene, not a photograph of a tested board
Embedded System Development · Nov 15, 2026

What Each ESP32 Family Does in a Product: Connectivity, Voice, Low-Power Nodes, and Displays

A Wi-Fi controller, a battery-powered Thread sensor, a voice terminal, and a camera-equipped display do not share one sensible ESP32 shortlist.

Read article
AI-generated scene of a wall-mounted control panel in a home, not a product or field test
AI and Machine Learning · Nov 11, 2026

Upgrading a Smart Control Panel for AI: Fix the Interaction Before Replacing the Board

A user says “get ready for bed” to a wall panel. It replies “done.” The living-room light may still be on, the wrong bedroom may have been selected.

Read article
Generated edge computing cabinet scene, not a measurement of a specific device
AI and Machine Learning · Nov 03, 2026

Can an 8 GB Edge Device Run a 2 GB Model? Budget Memory at Peak Concurrency

Eight gigabytes of RAM minus a 2 GB model file is not 6 GB of usable inference memory. Calculate KV cache, then test the longest request and peak concurrency.

Read article
Generated compute-module version-check illustration, not a customer batch or test photograph
Embedded System Development · Oct 31, 2026

Rockchip Compute Module RFQ: Align Revision, Carrier Board, and Software Image

Two suppliers quote an “RK3588 compute module.” One sample boots on the supplier's development carrier; the other quote is cheaper.

Read article
Generated edge-board development scene, not the target RK3588 device or a measured inference result
AI and Machine Learning · Oct 30, 2026

Deploying RKLLM on RK3588: Pin Versions, Convert the Model, and Validate On-Device Inference

Copying Hugging Face weights onto an RK3588 board does not make the model run on its NPU. Rockchip's RKLLM path has two distinct sides.

Read article
Generated module procurement scene for illustration, not an actual Tuya SKU, customer quote, or factory test
Embedded System Development · Oct 29, 2026

Buying Tuya Modules in Volume: Align Part Numbers, Firmware, and Production Tests

Two module quotes can both say “Wi-Fi” yet describe different products for your factory. One may need a different antenna layout.

Read article
AI-generated controller and inspection fixture; not a project photograph or test result
Embedded System Development · Oct 27, 2026

Adding Tuya to an Existing Product: What Changes Between Prototype and Production?

Start by assessing whether the existing controller and local control logic can stay. Then define the wireless module, firmware integration.

Read article
Generated edge LLM testing scene, not a measurement of any device
AI and Machine Learning · Oct 26, 2026

How to Benchmark an Edge LLM: TTFT, TPOT, Memory, Thermals, and Power

An edge LLM benchmark needs separate TTFT, TPOT, memory, thermal, and whole-system power results under a fixed model and workload.

Read article
Generated edge-computer selection scene, not a performance test of a specific board
AI and Machine Learning · Oct 23, 2026

How to Select Edge Hardware for LLM Inference: Memory, Bandwidth, Accelerators, and Power

A supplier's “8 GB RAM, 6 TOPS NPU” line does not establish that a field device can answer questions about a long service manual offline.

Read article
Talk to ZedIoT

Need to apply one of these ideas to your device or platform?

Share your device type, protocol, site conditions, platform goals, or AI workflow. We can help you evaluate the most practical implementation path.