/ blog / tag

#llm

11 posts

← all posts
01 Aug 19, 2026 16 min read

Testing Applications with an LLM Inside: How QA Handles Non-Determinism

How to test apps with an LLM inside: mocks for the deterministic code, structure contracts, property assertions, pass-rate evaluations and a calibrated LLM judge.

ai qa
Grzegorz Holak
02 Aug 13, 2026 9 min read

Synthetic Test Data with LLMs: How to Generate Realistic Data Without Leaking Production

A data contract, a deterministic validator, targeted edge cases, and the three traps of synthesis - how to generate test data with an LLM without copying production and without GDPR risk.

ai qa
Grzegorz Holak
03 Aug 6, 2026 9 min read

Hallucinations Under the Hood: Where They Come From and Five Techniques That Actually Reduce Them

Why a model makes things up by design rather than by malfunction, and five everyday defense techniques with before/after examples: forcing sources, a sanctioned "I don't know", task splitting, second-model verification, and grounding in documents.

ai hallucinations
Grzegorz Holak
04 Aug 5, 2026 9 min read

LLM as a Judge: Can a Model Evaluate Test Results, and When Not to Trust It

The LLM-as-judge pattern in testing: a rubric instead of a 1-10 scale, calibration on golden examples, three working modes, and four judge biases you need to know before rollout.

ai qa
Grzegorz Holak
05 Jul 30, 2026 9 min read

Fine-tuning, RAG, or a Longer Prompt? A Decision Tree Without the Marketing

A five-step decision tree for teams without an ML department: when a prompt with examples is enough, when you need RAG, and when fine-tuning genuinely pays off - with a table of cost, time, and risk.

ai rag
Grzegorz Holak
06 Jul 6, 2026 10 min read

AI in Smart Home, Part 5: A Home That Talks - A Voice Assistant with a Local LLM

The fifth part of the AI in smart home mini-series. We build a voice assistant on Home Assistant Assist with a local language model: the chain from wake word to speech synthesis, hardware choices, and an honest comparison with commercial assistants.

ai smart home
Julia Sielska & Grzegorz Holak
07 Jun 26, 2026 17 min read

Building Your Own AI Agents: From Personal Assistant to Specialist Team

How to build your own AI agents: from a single personal assistant to a team of specialized agents. A practical take on design, roles, and coordination, without magical thinking.

ai ai agents
Julia Sielska & Grzegorz Holak
08 Jun 12, 2026 9 min read

AI in Everyday Life: LLMs as Board Game Assistants

An LLM can be a very practical helper during board games: explaining rules, resolving ambiguities, searching rulebooks, creating turn summaries, and keeping house rules consistent. The condition is simple: treat it as an assistant, not an infallible judge.

ai llm
Julia Sielska & Grzegorz Holak
09 May 8, 2026 10 min read

Open WebUI - a ChatGPT-like frontend for your local LLM

Local Ollama only gives you a CLI. LM Studio is single-user. Open WebUI is the missing piece: a ChatGPT-like UI with RAG, web search and tools - running in one Docker command.

ai llm
Grzegorz Holak
10 May 7, 2026 11 min read

Local LLM models in 2026 - what actually runs on a Mac mini M4 16 GB

A review of the current models worth pulling onto 16 GB unified memory: gpt-oss-20b, Gemma 4 e4b, Qwen3-Coder, Phi-4. What works, what doesn't, and why.

ai llm
Grzegorz Holak
11 May 6, 2026 10 min read

Mac mini M4 16 GB as a local LLM workstation - LM Studio vs Ollama

Does the cheapest M4 Mac make sense as a machine for local language models? Hardware reality check, LM Studio vs Ollama comparison, and where the sanity line is.

ai llm
Grzegorz Holak