Hi, I'm Riddam.
Staff engineer in Amsterdam, working in cloud and application architecture. I build systems that hold up in production, read books worth remembering, and think about how to lead engineers well — and I write it all down here.
Latest
The AI Vocabulary: Seventy Terms That Unlock Everything Else
Seventy words you need before any other AI writing makes sense — token, context window, embedding, context engineering, tool call, MCP, agent, eval, prompt injection and the rest — each in one sentence with one concrete example, grouped by the thing you are trying to understand.
More writing
-
Skills, subagents, hooks, MCP, slash commands, plugins, scopes — what each part of an AI coding harness is actually for, organised by who pulls the trigger, and mapped across Claude Code and GitHub Copilot.
-
I checked which Claude Code plugins I actually reach for, and found one I was certain I used and never had. What a plugin really is, the integrations that let me find developer-experience gaps instead of guessing at them, why an idle plugin is usually a role mismatch rather than a bad one, and where to start.
-
A working process for junior developers in the AI era — taking a ticket from question to production, what to hand to the AI and what to keep, how to learn without letting the tool learn for you, and what reviewers actually look for.
-
What spec-driven development actually adds when an AI agent writes the code — one feature built three ways with TDD, BDD and SDD, how the three stack, and how to run the loop in Claude Code, GitHub Copilot, Cursor, Spec Kit and Kiro.
-
The Postgres schema that lets you re-extract a corpus with a new parser, compare both versions in SQL, promote atomically and roll back with one UPDATE — with every extracted value tracing back to a page and a bounding box.
-
What changes when AI work is throughput rather than conversation — the parser is the swappable seam instead of the model, queues replace serving, and residency gets enforced by queue topology rather than application code.
-
Hybrid AI platforms fail at the network boundary, not the model layer — outbound-only architecture, CIDR planning, choosing between Cilium Cluster Mesh, Istio and no mesh at all, and portable identity across the edge.
-
What it takes to lead engineers when you have no reports — earning a mandate, writing ADRs that land, building a bench of complementary working styles, spending your vetoes, and encoding influence into defaults.
Explore
- Engineering Deep dives and write-ups from real work — cloud and application architecture, and the reasoning behind the decisions.
- Book Notes What I took away from books worth remembering, in my own words.
- Leadership On leading teams, growing engineers, and doing the work well.
- Playbooks In-depth references and blueprints I've built for certifications, architecture, and core topics — the material I share with the engineers I work with.