patterns
Reads your rejected and discarded applications and surfaces patterns — JD shapes that score high but never convert, archetype mismatches, comp-band issues. The output is a calibration signal for your targeting filter.
What it does
The patterns mode reads every report in the rejected and discarded stages, looks for common features (industry, company size, seniority level, specific keywords), and produces a structured analysis of what is and is not working in your targeting.
When to use it
Run patterns after every 20–30 rejections to recalibrate. The mode surfaces blind spots in your profile and helps you avoid sending applications into the same dead patterns.
What a run looks like
You run the mode after a batch of outcomes and it returns a dated pattern-analysis report plus a short summary of the three findings that matter most. Under the hood it runs analyze-patterns.mjs over your tracker and reports, then reads back a conversion funnel, a score-versus-outcome table, top blockers, and a recommended score floor.
/career-ops patternsPattern Analysis Complete (24 applications, Apr 7 - May 18)
Key findings:
- Geo-restricted roles: 0% conversion (7 of 24) — stop evaluating US-only postings
- Regional/global remote roles convert at 57-67% — your sweet spot
- No positive outcomes below 4.2/5 — treat that as your score floor
Full report: reports/pattern-analysis-2026-05-18.mdAfter presenting the summary, the mode offers to act on the findings — updating portals.yml filters, setting a score threshold, or reweighting which archetypes you target.
When to reach for it
Reach for patterns once enough applications have moved past evaluation that outcomes carry signal. The hard gate is five entries with a status beyond "Evaluated" — Applied, Responded, Interview, Offer, Rejected, or Discarded; below that the mode exits and asks you to come back with more data. More outcomes mean sharper patterns, so it earns its keep after a month of steady applying.
- If you keep interview transcripts in
interview-prep/sessions/, the mode adds a targeting signal: where your most fluent, specific answers cluster, which catches when your strongest role-type differs from the one you keep applying to. - The report never names a real interviewer or company from those sessions — it summarizes competency clusters only, so anything committed stays private.
Related
tracker
Quick read of where every application stands. Useful as a daily check-in or when preparing the weekly review. The full keyboard-driven UI lives in the Go TUI dashboard; this mode is the chat-side summary.
followup
Flags applications and interviews that are overdue for followup based on the cadence rules you configure. Drafts the followup messages so you can review and send.