Technical work
These are the parts of my projects I spend time looking after. The notes cover decisions such as where to cache a page, how to store crawl results and what happens when a release fails.
Browse by topic
- 01AI and agentsmodels, tools, context, evaluation, approvals
- 02ApplicationsNext.js, React, Node.js, TypeScript, WordPress, PHP
- 03Searchcrawling, rendering, structured data, AI answer engines
- 04DataPostgreSQL, MariaDB, ClickHouse, Redis, pipelines, BI
- 05Analyticswarehouses, semantic layers, dashboards, AI analysts
- 06EdgeOpenResty and Lua, WAF, bot defence, caching, TLS
- 07OrchestrationNomad, Consul, Docker, Traefik, canary deploys
- 08Virtualisation and storageKVM, Proxmox, ZFS, RAID, iSCSI
- 09NetworkBGP, DNS, mail, firewalls, VLANs
- 10Hardwareservers, switches and power in a data-centre rack
Code review and AI
When I review a change to a page, I look at what it does to the cache and how often it queries the database. I also write about AI and engineering on LinkedIn, including my AI prediction for 2029.
In the G7 Agentic Platform prototype, reviewers see the version tested on staging. If that version fails after release, the monitor can roll it back.