Sairam S

Data engineer. I build maintained data feeds for the AI ecosystem: what changed, since when, from which source.

Every Actor I publish follows the same rules:

• Feeds, not snapshots. Each run answers "what is different since last time" and stores the baseline in the key-value store.
• Fail loudly. A source that breaks is reported as an error with a reason. No silent empty datasets.
• Machine-first output. Flat JSON records with stable keys, written for scripts and agents, not for humans reading a table.
• Usage-billed. Pay per event. A quiet run costs cents.
• Tested on real pages. Captured evidence for every release lives in the public repo.

Current Actor: LLM Pricing Change Feed , which watches 10 provider pricing pages and emits a diff per run.