Fit analysis#
Analyze this role on an application page produces a structured read of how you line up against that posting.
A real analysis: 84/100, skill-by-skill coverage, and the gaps worth closing.
The fit score#
A single 0–100 number, calibrated against explicit bands rather than left to the model's judgement:
| Range | Meaning |
|---|---|
| 85–100 | Would be shortlisted immediately; clears every hard requirement |
| 70–84 | Strong candidate with one or two soft gaps |
| 50–69 | Plausible, but the gaps need addressing to get past screening |
| 25–49 | A stretch; needs a referral or a manager willing to train |
| 0–24 | Not a credible match |
The prompt tells the model to score like a hiring manager, not a cheerleader, because an inflated score costs you interviews — you spend effort on the wrong roles and skip preparation you needed.
What else you get#
Skill coverage — every significant skill the posting asks for, each marked have it, partial or gap, with a clause of evidence. These use the reserved status colours and always carry a text label, so the state never depends on colour alone.
Lead with these — three to five specific things to foreground in this application.
Close these gaps — the honest weaknesses. These are what the prep plan targets.
Likely interview focus — the topics this company's loop will probe.
Compensation read — a realistic band for this role, market and level, compared against the posted range and your target.
How to position yourself — concrete résumé and cover-letter angle for this specific role.
Re-running#
Re-run produces a fresh analysis and keeps the old one in the timeline, so you can see how the read changed after you added the full job description or updated your résumé.
What drives quality#
- Your résumé in Settings. Analysis is a comparison; with nothing to compare against, Alfred says so rather than inventing a background for you.
- The job description. Without it, Alfred reasons from the title and seniority and flags the uncertainty.
- The model. A frontier model reasons about skill adjacency — that your Postgres partitioning work is evidence for a "data modelling at scale" requirement. A small local model tends to match keywords.
Alfred will not invent experience
The prompts forbid attributing experience you haven't stated. If the analysis looks thin, the fix is a fuller résumé, not a different question.