Filters are hard constraints. If a candidate does not satisfy a filter, they are removed before final ranking. Use filters for non-negotiables, not preferences.

MatchFilters field reference

MatchFilters

Field	Type	Required	Default	Constraints	Notes
skills	string[]	No	—	—	Required technical skills.
languages	string[]	No	—	—	Required spoken languages.
country	string	No	—	—	Country hard constraint.
city	string	No	—	—	City hard constraint.

How to design filters safely

Start broad, then tighten

Begin with role description + top_n only. Add one filter at a time and compare outcome changes.

Use filters for mandatory constraints

Use country if legal/work-permit constraints require it. Do not use location filters for "nice to have" preferences.

Avoid long skill hard-filters

Treat long skill lists as soft guidance in the job description whenever possible. Hard filters with many skills often over-constrain results.

Practical examples

{
  "job_description": "Cloud platform engineer",
  "filters": {
    "country": "Belgium"
  }
}

{
  "job_description": "Data engineer with strong SQL and ETL ownership",
  "filters": {
    "skills": ["SQL", "ETL"],
    "languages": ["English"]
  }
}

Why expected candidates disappear

Filters are stricter than intended
Job description is too narrow and duplicates filter constraints
Candidate profiles are missing data for filtered fields

Debug pattern

Remove filters first. If candidate appears, re-add filters one by one to identify which field is over-constraining your query.