Searcher: each track independently tracks when it was last searched
via data/track_history.json. New tracks get full lookback (90d),
existing tracks look back since their last completion. Keyword-level
crash resume preserved.
Profiles: search tracks can specify profile_overrides (inline) or
profile_path (external file) to customize resume, cover letter, and
experience per track. Filter and applier both use the track-specific
profile. Base profile.json provides shared info (name, contact, etc).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Search config already defines what each track is looking for (keywords,
exclude_keywords, salary_min, remote). Profile.json defines who the
candidate is. No need for a third file duplicating both.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- filter.mjs: loadProfile now async, uses loadJSON
- telegram_answers.mjs: answers read/write through storage layer
- status.mjs: uses initQueue + loadQueue for S3 support
- setup.mjs: await all loadConfig calls
- storage.mjs: more robust getS3Key using URL parsing
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- submitBatch → submitBatches: groups jobs by track, submits one batch each
- filter_state.json now stores batches[] array instead of single batch_id
- Collect waits for all batches to finish before processing
- Each track gets its own cached system prompt = better caching + cleaner scoring
- Idempotent collect: skips already-scored jobs
- Batch API = 50% cost savings vs synchronous calls
- Prompt caching on system prompt (profile + criteria shared across all jobs)
- One request per job with custom_id = job ID for result matching
- Two-phase state machine: submit → poll/collect (hourly cron safe)
- filter_state.json tracks pending batch ID between runs
- Model configurable via settings.filter.model (default: claude-sonnet-4-6)
- Telegram notifications on submit + collect
- Errors pass through — never block applications due to filter failure
- --stats flag for queue overview
- lib/filter.mjs: batch scoring engine (10 jobs/call, Claude Haiku)
- job_filter.mjs: standalone CLI with --dry-run and --stats flags
- Threshold configurable globally + per-search in search_config.json (filter_min_score, default 5)
- Job profiles (gtm/ae) passed as context via settings.filter.job_profiles
- Filtered jobs get status='filtered' with filter_score + filter_reason
- Filter errors pass jobs through (never block applications)
- status.mjs: added 'AI filtered' line to report
