Field Notes, Architecture &
Practical AI Writing.
Essays written for busy technology leaders and engineers. No marketing fluff — only actionable frameworks, tested patterns, and honest lessons from production code.
AI for Busy Professionals: The Useful Mental Model
Why treating LLMs as probabilistic interns rather than infallible experts gives you predictable leverage without suffering costly hallucinations.
LLMs Demystified: Tokens, Context Windows, and RAG
A plain-English explanation of how next-token prediction works, where context limits bite, and why Retrieval-Augmented Generation is essential for accuracy.
Working with Coding Agents: Claude Code, OpenCode, and Clear Guardrails
How to structure terminal coding harnesses, enforce deterministic tests, sandbox file access, and avoid catastrophic code rewrites.
From Repetitive Tasks to Reliable Pipelines: Practical AI Automation
The 4-stage framework that turns noisy manual operations into resilient pipelines with structured JSON schemas and human-in-the-loop review gates.
AI in Financial Trading: Research, Backtesting, and Why Risk Management Always Wins
Using machine intelligence for market research while keeping mathematical backtesting and hard circuit breakers strictly non-negotiable.
Have feedback or want to discuss an article?
Reach out directly to James on LinkedIn for engineering discussions.