How Sift scores every job against your profile

Every job in Sift gets a score from 0 to 100. The number is meant to answer one practical question: is this role worth your time?

A score is not a guarantee, and it is not a judgment of your ability. It is a fit signal. It helps you separate roles that match your profile from roles that only look interesting because the title is familiar or the company is exciting.

What the AI reads

Sift reads the full job description: title, responsibilities, requirements, preferred qualifications, company information, location, and compensation when the range is listed. It looks for explicit signals, like a required programming language, and softer signals, like whether the role is closer to founder-led sales, enterprise account management, product marketing, or technical customer work.

It also reads your profile. That includes your skills, experience level, target roles, location preferences, compensation range, and industry preferences. The goal is not to create a generic "good job" score. It is to score the role against you.

That distinction matters. A senior backend engineering role at a great company might be an excellent job and still be a poor match for someone targeting B2B SaaS sales. Sift is designed to make that mismatch visible before you spend time tailoring an application.

The scoring factors

The composite score is built from several factors. Skills alignment asks whether your skills match what the role needs. Experience level checks whether you are applying at the right seniority. Location fit compares remote, hybrid, and on-site requirements against your preference. Compensation checks whether the listed range fits your target. Industry and company fit look at whether the organization is the kind of place you said you want to work.

Each factor contributes to the final score. A 91 means strong alignment across most dimensions. You likely have the skills, the level is right, the location works, and the company type makes sense. A 42 means there is a significant mismatch somewhere. Maybe the role is too senior, too junior, outside your target geography, or built around skills you do not want to use.

The point is not to let the score make every decision for you. The point is to make tradeoffs visible. If a role scores 58 because the location is imperfect but the company is exactly what you want, you can still choose to apply. You are just choosing with context.

How to use scores

Focus your time on roles scoring 70 or above. These are the jobs worth reading closely, tailoring a resume for, and tracking with care. They are not automatic yeses, but they are good candidates for serious attention.

Below 50, the mismatch is usually too large. Wrong level, wrong location, wrong industry, wrong skill mix. You can still review them, but the score is a warning that your energy may be better spent elsewhere.

Between 50 and 70, skim the description and decide. These are the edge cases: interesting companies, unusual roles, or postings that may be missing enough detail for a confident score. Sometimes the best opportunities live there. Sometimes they are simply maybes.

The model

Sift uses Claude Haiku 4.5 for scoring: fast, accurate, and cost-effective. Each score costs a fraction of a cent, which means you can scan enough jobs to find signal without turning matching into a luxury feature.

The free tier includes 50 scores per month. That is enough to test the workflow, build a shortlist, and see how different roles rank against your profile before you spend hours rewriting resumes.

Your top matches are waiting. Start with Sift.