PostgreSQL internals, properly
Go from writing SQL to explaining what the planner, storage engine and MVCC are doing underneath it. Eight-week self-directed course with weekly assignments.
Jump to a page, project or post
Roadmap
Grouped by theme and status rather than pinned to dates. Timeframes are deliberately loose; the statuses are kept honest. Shipped items link to their project or progress page.
Updated 27 Oct 2026
Go from writing SQL to explaining what the planner, storage engine and MVCC are doing underneath it. Eight-week self-directed course with weekly assignments.
A nine-phase build-it-yourself pipeline in Rust and rdkafka: logs and offsets, producers, consumer groups, delivery semantics, schemas, a three-broker cluster, stream processing, performance tuning and failure drills.
Ship the pilot, then refine over a few days: design passes, resume PDF export, OG images per page, GitHub activity on the progress page, and a proper domain.
Provision real infrastructure from scratch: VPCs, bastions, ECS Fargate behind an ALB, then rebuild all of it with Terraform and wire up CI/CD. Certifications along the way: AWS SAA-C03, then DVA-C02 or SOA-C02, then CKA.
An assistant in the corner of every page that answers questions about my work, resume, roadmap and progress, grounded in the same content files that render the site.
Build a small three-node system with a queue, a database and a cache, then break it on purpose: partitions, clock skew, slow disks, split brain. Write up what each failure looks like from the outside.
Land a non-trivial change in a crate I depend on at work, most likely in the Kafka or async-database space, and write up the review process.