Career Pipeline Workflow
Daily career operations scan roles, evaluate fit, update pipeline recommendations, and flag resume refresh work without auto-applying or rewriting resume source.
The career-pipeline automation keeps Career Profile, Resume, Kevin B. Liu, and the job-search operating loop aligned with the live market. It is recommendation-first: agents surface roles, scores, and resume gaps; Kevin decides applications and resume changes. Source: automations/career-pipeline.md, 2026-05-31
Routing
Run this workflow for daily job-market maintenance, not for one-off resume rewriting. Use Recruiter Discovery Workflow for recruiter relationship discovery and Application Standout Protocol when a specific application needs a full package.
Trigger
- Daily cloud cron from
automations/career-pipeline.md. - Manual request: "Run the
career-pipelineautomation." - Follow-up after new Dedalus accomplishments, new target-company direction, or a known hiring-market shift.
Inputs
- Target company tiers, role archetypes, and career constraints from Career Profile.
- Current resume and resume-source state from Resume, Kevin B. Liu and the LaTeX source when present.
- Greenhouse, Ashby, Lever, Wellfound, and the career-ops scanner.
- Existing application tracker or the last workflow output when available.
Runbook
- Load Career Profile and identify the current company and role filters.
- Scan portals for founding engineer, SWE, SDE, ML engineer, and AI engineer postings.
- Deduplicate by company, role title, location, seniority, and posting URL.
- Run career-ops archetype detection and quick A-F evaluation.
- Flag roles scoring 4.0+ for full evaluation, keep 3.0-3.9 as watchlist, and reject below 3.0 unless Kevin explicitly requested broad exploration.
- Compare high-scoring postings against current resume bullets and recent Dedalus accomplishments.
- Report new roles, expired roles, follow-ups, and resume deltas.
Promotion
Update durable wiki pages only when a finding changes the operating state:
- Update Career Profile when target companies, role filters, compensation constraints, location constraints, or search strategy changed.
- Update Resume, Kevin B. Liu when there is a real resume gap or approved bullet direction.
- Append timeline entries for decisions and material pipeline movement.
- Do not auto-apply, auto-edit LaTeX, or mark a role as applied without Kevin's explicit confirmation.
Output
Write the run result to outputs/<YYYY-MM-DD>/career-pipeline/<agent>.md when run as an automation. The output should include new roles, full evaluations for 4.0+ roles, resume suggestions, rejected-role rationale, and pipeline status. If no roles qualify, record what sources were checked and why the run is a no-op.
Validation
A good run can point to source URLs, scoring rationale, and the specific career pages that did or did not change. The run is incomplete if it has recommendations without source links, scores without criteria, or state updates without an output artifact.
Run Contract
This workflow follows Workflow Run Contract: name the sources, write output or no-op proof, promote only durable facts, update state only when the run really completed, refresh generated surfaces when durable pages or skills change, and log user-visible work.
Timeline
- 2026-07-01 | Expanded from a thin automation stub into a runnable career operations contract with routing, trigger, inputs, promotion rules, output, validation, and human-gated side effects. Source: User request, 2026-07-01
- 2026-05-31 | Workflow page created from career-pipeline automation. Source: automations/career-pipeline.md, 2026-05-31