OH SHI
A public source of truth for startup companies, hiring signals, canonical job status, and recent changes—built for people and AI agents.
Startup hiring, with receipts
OH SHI is the Operational Headquarters for Startup Hiring Intelligence. It gives people a fast, searchable job board while giving agents stable JSON, JSONL, incremental change feeds, evidence URLs, and explicit open or closed states.
The product focuses on a common failure in job data: listings are copied, reposted, and left online long after the canonical role closes. OH SHI checks supported company hiring sources directly and records what changed.
The current product
- Search and filter verified-open roles across tracked startups.
- Inspect company-level hiring momentum, recent changes, source freshness, and evidence confidence.
- Read the same ranked inventory through human pages and versioned agent interfaces.
- Follow canonical ATS links rather than storing or presenting a repost as the source of truth.
- Track openings, closures, company additions, and dated funding signals as separate events.
Human and agent parity
The public interface and agent endpoints share the same deterministic ordering and status model. The preferred intelligence endpoint advertises its capabilities, validates filters, supports cursor pagination, and exposes natural-language request recipes through llms.txt.
That parity matters: an agent should not see a looser or less explainable version of the data than a person does.
Trust model
Hiring momentum and evidence confidence are separate, deterministic scores. Neither is presented as a probability of being hired. Records link back to canonical evidence, stale sources decay, and incomplete or failed refreshes do not fabricate certainty.
OH SHI is live and refreshing its public hiring inventory while coverage, source adapters, and audit tooling continue to expand.