Pipeline Sitter
Watches ETL jobs and retries the flaky ones with backoff before you wake up. Built for Quiet Hours — build an agent that does useful work while its user is asleep. Retry orchestration and the morning digest. The interface is deliberately narrow: one path, done properly, rather than a broad surface that half-works at the deadline.
Gallery
Screens
Pipeline Sitter — main view
Pipeline Sitter — result state
On the clock
What shipped, what didn't
Completed in time
Retry orchestration and the morning digest.
Unfinished / faked
No distinction between flaky and genuinely broken.
Obstacles
Rate limits hit at minute 20 and forced a caching layer I hadn't planned for.
Lessons learned
Hardcode the thing you can hardcode. Nobody is scoring your config system.
Process
Prompting approach
Started from a single system prompt describing the user and their constraint, then iterated with the assistant on the Pipeline Sitter data model before writing any interface code. Full prompt log is in the repository under /prompts.
Fairness
Disclosures
Declared by the builder at submission time, visible to judges before scoring.
Judging
Public feedback
Judges' private notes stay private — this is what they chose to say publicly.
Elena Voss
VP of Engineering, Harbor Compute
This needed a working demo path. I could not verify the core claim from what was submitted.
Hana Kobayashi
Principal Research Engineer, Independent
The concept is interesting but not enough shipped to evaluate it fairly. Scope down harder next time — pick one screen and finish it.
Claudia Reyes
Staff Engineer, Vantage Analytics
The concept is interesting but not enough shipped to evaluate it fairly. Scope down harder next time — pick one screen and finish it.
Discussion
0 comments
No comments yet.