Skip to main content

​
Skills & Technology Filters

Skills filtering is one of the most powerful features in Client Discovery. It lets you find companies based on the actual technologies and skills mentioned in their job postings, giving you a direct line to companies that need the type of candidates you place. While industry tells you what sector a company is in, skills tell you exactly what kind of talent they need right now.

​
How Skill Filtering Works

Unlike industry data that comes from LinkedIn company profiles, skill data is extracted from job postings. Here is the process:
  1. Job postings are scraped from LinkedIn and other sources.
  2. Skills are extracted from each job description using natural language processing. The extraction analyzes the full job description text and identifies specific technologies, tools, programming languages, frameworks, and professional competencies.
  3. Skills are clustered into base skills and variants to handle naming differences (e.g., “ReactJS”, “React.js”, and “React” are all grouped together under the same base skill).
  4. Skills are stored in a structured format as part of the job’s requirements_summary field, which includes both the skill names and their context within the job description.
  5. Skills are linked to companies through their job postings, creating a technology profile for each company based on what they are actually hiring for.
When you filter by a skill like “Python”, Recruitier finds all companies that have at least one active job posting mentioning Python. The search goes beyond the company profile to look at what the company actually needs in its current hires.
Skills data is only as current as the active job postings. If a company stopped posting Python jobs but still uses Python internally, they would not appear in a Python skill filter search. This is actually a feature, not a limitation — you want to find companies that are currently hiring for specific skills, not just companies that use those skills internally.

​
The Hierarchical Skill System

Skills in Recruitier are organized in a two-level hierarchy:

​
Base Skills

These are the top-level skill categories. For example:
  • React
  • Python
  • Java
  • AWS
  • Docker
  • SQL
  • Kubernetes
  • TypeScript

​
Skill Variants

Each base skill can have variants that represent more specific technologies or combinations. For example, the base skill “React” might have variants like:
  • React/Frontend
  • React/Next.js
  • React/TypeScript
When you select a base skill, you match jobs containing any of its variants. You can also select specific variants if you want to be more precise about the exact technology combination.
The skill hierarchy is built automatically from the job data. Variants are created when skills frequently co-occur in the same job postings. A variant like “React/Next.js” means jobs that mention both React and Next.js together, indicating a specific technology stack rather than a general React requirement.

​
Searching by Technology Stack

The skill filter provides a searchable dropdown where you can:
  1. Type to search — Start typing a skill name and the dropdown filters to matching skills instantly. This works for both base skills and variants.
  2. See job counts — Each skill shows the number of active jobs requiring it, helping you gauge market demand at a glance.
  3. Select multiple skills — Add as many skills as you want to build a complete technology profile for your target companies.

​
Skill Match Modes

When you select multiple skills, you can choose how they are combined:

Match Any (OR Mode)

Default mode. A company appears in results if it has jobs requiring any one of your selected skills. Use this when you are casting a wide net across related technologies.Example: Selecting “React”, “Angular”, and “Vue.js” finds companies hiring for any frontend framework. This is ideal when you place frontend developers who might know any of these technologies.

Match All (AND Mode)

A company only appears in results if it has jobs requiring all of your selected skills. Use this when you want companies with a specific technology combination.Example: Selecting “Python” and “AWS” in AND mode finds companies that use both Python and AWS, not just one or the other. This targets companies with a specific cloud-based Python stack.
Start with Match Any mode to see a broader set of results, then switch to Match All if you need to narrow down to companies using a specific combination of technologies. You can toggle between modes without losing your skill selections.
In Match All mode, adding more skills progressively narrows results. If you select 5 skills in AND mode, a company must have jobs mentioning every single one. This can quickly reduce results to zero if the combination is too specific. Use AND mode with 2-3 complementary skills for best results.

​
Skill Variants in Practice

Understanding skill variants helps you search more effectively:
Base SkillCommon VariantsWhat This Means
PythonPython/Django, Python/FastAPI, Python/Data ScienceDifferent Python ecosystems
ReactReact/Frontend, React/Next.js, React/TypeScriptDifferent React stacks
AWSAWS/DevOps, AWS/Cloud Architecture, AWS/LambdaDifferent AWS specializations
JavaJava/Spring, Java/Microservices, Java/BackendDifferent Java contexts
DockerDocker/Kubernetes, Docker/CI-CD, Docker/DevOpsDifferent containerization contexts
When you select a base skill, all its variants are included. This means selecting “Python” will find companies posting Python/Django jobs, Python/FastAPI jobs, and general Python jobs alike.

​
Common Use Cases

​
Finding Companies That Need Your Candidates

If you have a pool of React developers, search for companies posting React jobs:
1

Add React to Skill Filters

Select “React” from the skills dropdown. This includes all React variants (React/Frontend, React/Next.js, etc.).
2

Optionally Add Related Skills

If your candidates also know TypeScript or Next.js, add those too in Match Any mode. This broadens the search to include companies using any of these related technologies.
3

Combine with Location

Set a location center and radius to focus on your geographic market. This ensures the companies you find have jobs your candidates can actually reach.
4

Sort by Hotness Score

Companies hiring most intensively for these skills appear at the top. A company with a high hotness score and matching skill requirements is a prime outreach target.

​
Identifying Companies by Technology Stack

If you specialize in placing DevOps engineers, find companies with a modern DevOps stack:
  1. Select skills like Docker, Kubernetes, AWS, Terraform, and CI/CD.
  2. Use Match Any mode to find companies using any of these tools.
  3. Review the results to identify companies with the deepest DevOps needs (highest matching job counts).
Look at the matching vs. total job count for each company. A company showing “8 / 10” matching jobs has DevOps as a core hiring need. A company showing “1 / 50” has one DevOps role among many other unrelated positions. Both might be worth pursuing, but the first is a much stronger alignment with your specialization.

​
Understanding Market Demand

The skill filter also serves as a market intelligence tool. The job counts next to each skill show you:
  • Which technologies are most in demand — Skills with high job counts indicate strong market demand and potentially competitive placement fees.
  • Emerging technologies — New skills appearing in the list may signal growing market trends that you can position yourself for early.
  • Niche opportunities — Skills with lower job counts but specific company interest may represent underserved niches where fewer recruiters are competing.

​
Combining Skills with Other Filters

Skills filters combine with all other filters using AND logic:
CombinationResult
Skills + IndustryFind IT companies that specifically need Python developers
Skills + LocationFind companies near Amsterdam posting Java jobs
Skills + Employee CountFind startups (1-50 employees) using React
Skills + Hiring ActivityFind companies that posted Kubernetes jobs in the last 7 days
Skills + Hotness Score SortFind companies hiring most aggressively for your specific skills
The skill filter only shows skills that appear in a minimum number of job postings (typically 5 or more). Very niche or rarely mentioned skills may not appear in the dropdown but could still exist in individual job descriptions. If you are looking for a very specific skill that does not appear, try a broader related skill.

​
Tips for Effective Skill Searching

  • Use base skills for broad searches — Selecting “Python” captures all Python-related jobs regardless of framework.
  • Use variants for targeted searches — If you specifically place Django developers, selecting “Python/Django” gives you more precise results.
  • Check the job count — A company with 5 matching jobs out of 20 total has a stronger need for your specialization than one with 1 out of 50.
  • Combine with hotness score sorting — This surfaces companies where these skills represent a significant portion of their hiring needs.
  • Monitor skill trends over time — If a skill’s job count is growing, that sector is expanding and likely needs more recruitment support.

​
Advanced

​
How Skill Matching Works Under the Hood

Skill filtering operates on the requirements_summary JSONB field stored on each scraped job record. The requirements_summary contains a structured extraction of skills from the job description, organized by category. When you search for a skill, the system performs a JSONB substring match against the skills key within this field. The matching supports compound skill names like “React/Frontend” — these are matched as substring patterns within the JSONB data. This means the search is flexible enough to capture both exact matches and partial matches within compound skill identifiers.

​
Match Mode Query Logic

The two match modes translate to different SQL logic:
  • Match Any (OR): The query checks if the job’s requirements_summary->'skills' contains a match for ANY of the selected skill patterns. A company appears in results if it has at least one active job matching at least one selected skill.
  • Match All (AND): The query checks if the job’s requirements_summary->'skills' contains matches for ALL selected skill patterns simultaneously within the same job, or across different jobs at the same company. A company needs coverage across all selected skills.

​
How Skills Affect the Hotness Score

When skill filters are active, the hotness score calculation switches from using the total job count to using the matching job count. This means:
  • Without skill filters: hotness = total_jobs / employees
  • With skill filters: hotness = matching_jobs / employees
This switch is important because it changes the ranking. A company with 50 total jobs but only 2 matching jobs will have a lower hotness score (based on 2) than a company with 10 total jobs and 8 matching (based on 8), even though the first company has more total openings. The ranking adapts to prioritize companies where your specific skills represent a significant hiring need.

​
Connection to Other Features

  • Company Detail Page Skills Overview: When you visit a company detail page, the skills overview shows the top technologies across all of that company’s active jobs. This uses the same requirements_summary data that powers the skill filter.
  • Job Classification: Internally, jobs are also classified by function using ML-based title matching. Skills represent specific technologies, while job classifications represent broader role categories (Engineering, Sales, Marketing, etc.).
  • Saved Searches: Your selected skills and match mode are preserved in saved searches. When notifications re-run a saved search, they use the same skill filters.

​
Power User Tips

  • Compound skills reveal specializations: A skill variant like “Python/Data Science” indicates jobs that specifically combine Python with data science work. Selecting this variant instead of just “Python” targets a much more specific subset of companies.
  • Use OR mode for discovery, AND mode for targeting: Start with Match Any to understand the landscape, then switch to Match All when you know exactly what technology combination your ideal client uses.
  • Skills + Recency is a powerful combination: Filtering for a specific skill posted within the last 48 hours gives you companies that just started looking for that type of candidate — the perfect timing for outreach.
  • Empty skill results do not mean zero demand: If a very niche skill returns no results, it may be below the minimum threshold for the dropdown. The actual demand may exist in individual job descriptions that were not clustered into a named skill.

​
Business Logic Rules

  • Skills are extracted from job descriptions via NLP during the scraping process and stored in the requirements_summary JSONB field.
  • The minimum job count for a skill to appear in the dropdown is configurable (typically 5+ active jobs).
  • Skill matching is case-insensitive and uses substring patterns.
  • When skills filters are active, the matching job count subquery is used for both the displayed job count and the hotness score calculation.
  • Compound skills (e.g., “React/TypeScript”) are single entries in the skill hierarchy, not combinations of two separate skills.