AI and Machine Learning articles

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
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
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
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
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
Choosing an Edge LLM Runtime: Start with Devices and Model Artifacts
Compare llama.cpp, RKLLM, ExecuTorch, and MLC LLM by target device, model artifact, acceleration backend, and release responsibility—not an unverified speed ranking.
Read article
Offline RAG for Equipment Manuals: Versioned Evidence Before Answers
A technician asks how to reset an E42 alarm. A local model produces a convincing procedure, but cannot identify the machine variant, manual edition, or page it used.
Read article
When an RK3588 Edge AI Box Fits Industrial Vision—and When It Does Not
A 6-TOPS RK3588 does not automatically deliver a production vision system. Learn how to qualify camera, RKNN, latency, thermal, recovery, and rollback boundaries.
Read article
How to Choose a Local AI Model for Enterprise Deployment
The most expensive mistake in local AI selection is not choosing the second-best model. It is downloading a popular model before defining where data may go.
Read article
FunASR for AI Voice Projects: Offline, Edge, or Cloud ASR?
FunASR belongs on the shortlist when audio must remain in a factory, clinic, meeting room, or private network; transcription must continue during an Internet outage.
Read article
n8n AI Workflow Automation for Business Systems
n8n is most useful when it turns hand-offs scattered across webhooks, CRM, email, forms, tickets, and internal APIs into a visible process that can be replayed and.
Read article
YOLO for Vision Inspection: From Acceptance Criteria to Edge Deployment
YOLO is a strong fit for a specific class of vision problem: determine whether a visible object is present, identify its class.
Read articleEvaluating a AI and Machine Learning project?
Share the device, data flow, network condition, target users, and business system boundaries. We can help you decide what to build first.