Writing a brief
The brief is natural language — who to find, what signals matter, optional count target.
Write the brief the way you'd describe the role to a colleague — not a database query.
The components
Who — role, seniority, specialization: "Senior ML engineer", "VP of Sales", "Founding designer"
Where — city, region, or remote: "in Berlin", "London or remote", "DACH region"
Signals — what matters for this search: "fintech background", "open-source contributor", "sold to enterprises"
Count target — include a number to set a target; Orbit stops when it reaches it:
Senior software engineers with payments experience in Berlin — find 25Examples
CTOs and VPs of Engineering at Series A–C SaaS startups in London,
fintech or devtools backgroundFounding product designers at consumer mobile startups in NYC,
ex-Airbnb or ex-Spotify preferredGo engineers with distributed systems experience, remote OK,
open-source contributorsIteration
Orbit evaluates every result and assigns a verdict with a reason. Use the verdicts to refine:
- Too many
irrelevant→ add a disqualifier: "not recruiting agencies", "must have shipped production code" - Too many
unknown→ add clarifying signals: "must be currently employed", "fintech not crypto" - Too few results → broaden: drop a constraint, expand geography, add
orbetween roles
Clarification questions
If Orbit isn't confident, it pauses the project (status: "awaiting_input", clarificationQuestion populated). Answer in the Orbit UI — the search resumes automatically.