TORSDAG
2026-06-11

Too many projects, too many ideas, too few hours — one learning a day anyway

Projects

The software I build and run. Only the load-bearing ones are listed — there are always more experiments in the drawer. Each card opens a dossier: what it is, how it’s built, and every post it has generated. The home, the network and the boat live in The Lab.

A personal sommelier that lives on the phone, not in the cloud.

A wine app for iOS and Android that inverts the usual architecture: the intelligence — taste profiling, recommendations, scanning a restaurant wine list and picking the bottle — runs on-device. The backend stays deliberately thin: it syncs user state and ships a static catalog, and the heavy lifting (scraping shops, enriching data) happens in batch on local hardware. The interesting bet: modern phones are smart enough that the cloud-brain + dumb-client split is optional.

ARCHITECTURE local batch scraping + enrichment → static catalog in object storage → on-device scan / match / recommend → thin API for user-state sync
Swift · Kotlin · on-device AI · FastAPI 3 posts →

An autonomous AI dev team that turns GitHub issues into pull requests.

Label an issue, and a pipeline of AI agents picks it up: a project manager enriches it into a plan, a tech lead dispatches tasks, workers implement and open PRs. Humans review at gates. It’s an ongoing experiment in how much of the routine engineering loop can run unattended — and in what kinds of supervision structures make that trustworthy rather than terrifying.

ARCHITECTURE labeled GitHub issue → PM agent plans → tech-lead dispatches → worker agents implement → PR → human review gates
Python · Docker · multi-agent

Dashboards and workers for moving thousands of users between Microsoft 365 tenants.

Enterprise consulting work: tooling around a large Microsoft 365 tenant-to-tenant migration — a real-time status dashboard plus PowerShell workers orchestrating the migration platform, target tenant, and source directory. The interesting part is the shape of the problem: identity, mail, devices and shared services all moving while the business keeps running.

ARCHITECTURE migration platform APIs → PowerShell workers → orchestrator + dashboard (Flask/Vue) → live status for the cutover team
Flask · Vue · PowerShell · Cloud Run

This site — one learning a day, news on merit, co-written with an AI editor.

The site you’re reading. Learnings publish exactly one per day from a queue — backlog is runway, not debt. News posts flow on merit, fact-checked against fetched sources, my take on top. An AI co-writes and argues with me (its objections get surfaced, not silently applied), but every post passes a human gate. The system’s most important skill is saying no.

The full architecture drawing — doors, agents, the human gate, four lanes — lives at The Machine, public on purpose.

ARCHITECTURE ingest doors (fetch/news/today/agents) → co-writing session → human-gated queue → one/day publisher → static broadsheet
Astro · Docker · claude 1 post →

An always-on local AI orchestrator with long-term memory.

The operating environment behind everything else: a resident orchestrator that routes tasks to LLM workers, runs scheduled agents (like the weekly Field Monitor that feeds this site’s news), and maintains long-term memory — nightly ingestion of work sessions and notes into a searchable store. The design principle: the orchestrator plans and verifies; disposable workers execute.

ARCHITECTURE always-on orchestrator → routes tasks to LLM workers → scheduled agents (news, code-milestones) → long-term memory ingested nightly
LLM routing · cron agents · vector memory