ARTIFICIAL INTELLIGENCE2 min read
Running Local AI Without Heating the Entire House
What model size, quantization, and power measurements reveal about a genuinely useful local inference stack.
TECHNICAL JOURNAL · BUDAPEST
Field notes from
a curious system.
Field notes on local AI and inference, reliable DevOps pipelines and observability, Raspberry Pi builds, calm homelab networking, and the systems behind smooth gaming.
LAB / NODE MAP ALL SYSTEMS NOMINAL
04ACTIVE
NODES
NODES
19WIDLE
DRAW
DRAW
01 / FEATURED FIELD NOTE
LONG-FORM · TESTED IN THE LAB
02 / RECENT TRANSMISSIONS
Notes from the lab
DEVOPS
Building Fast Serial CLI Tools in Rust
Serial debugging is still central in embedded work, and Rust is excellent for building reliable terminal tooling. I structure serial utilities as pipelines: read bytes, frame messages, parse the protocol, then route structured events out.
DEVOPS
Rust no_std Patterns for Embedded Systems
Writing `no_std` Rust firmware requires different habits than server Rust. Memory is constrained, timing is strict, and panic strategy must be deliberate.
DEVOPS
Testing Hardware Abstractions in Rust
Hardware abstraction layers are where embedded code often becomes difficult to test. I design traits around capabilities, then keep business logic independent from concrete drivers.
DEVOPS
Rust Testing Pyramid for Hardware-Connected Projects
Hardware-connected software fails at boundaries: timing edges, flaky links, and rare protocol states. A strong test strategy must combine fast feedback with realistic integration checks.
03 / TOPICS
FIVE THREADS
Artificial Intelligence2 notesPractical AI: local models, inference, tooling, and honest measurement.Raspberry Pi17 notesSmall single-board computers put to real, tested use.DevOps17 notesBoring, reliable delivery: pipelines, observability, and operations.Homelab and Networking26 notesHome infrastructure and networking designed to be observable and calm.Gaming and Game Systems0 notesThe systems behind play: latency, hosting, state, and hardware.
04 / ACTIVE SERIES
ORDERED READING
Practical Local AI
Running useful AI on your own hardware without pretending the trade-offs do not exist.
Follow the series05 / THE HUMAN IN THE LOOP
PORTRAIT / ADOBE STOCK OR ORIGINAL
I build systems to understand them — then write down what the diagrams leave out.
I'm Gábor Lepsényi, a technologist and incurable tinkerer based in Budapest. I write about running local AI and inference, keeping DevOps pipelines and observability boring, Raspberry Pi and embedded builds, calm homelab networking, and the systems behind smooth gaming.
No recycled release notes and no ten-minute miracle builds — just careful experiments, honest constraints, useful measurements, and enough context to reproduce the work. You can also find me writing in Hungarian at rust.hu and aws.hu.