A personal notebook

Writing

I write about engineering with AI agents and building reliable agent-based systems: clear specifications, context, tools, orchestration, evaluations, observable execution and permissions. These are short notes and experiments about what I tried, what I measured and what remains uncertain. Ideas about future learning mechanisms are research questions, not demonstrated capabilities.

Topics

2 published

Agent systems

0

Orchestration, context, memory and the machinery around the model.

Safety & reliability

0

Permissions, containment, evaluation and failure handling.

Ideas & questions

0

What the evidence supports, what remains uncertain and what I want to test next.

All writing

Newest first

Jev vs LLM-as-a-Judge: What I Found Testing Both on a Real Agent

I ran our hand-written eval rules, TypeSafe's Jev and four frontier LLM judges side by side on 239 labelled cases from a real family-assistant agent. Here's where each one earns its place.

evalsllm-as-a-judgejevlangsmith
9 min read

BERT, an LLM, or Jev: what should make the decision?

Compare encoder classification, causal generation and Jev's bounded decision interface, while keeping its unknown internals explicit.

BERTJevAI engineering
17 min read