> ## Documentation Index
> Fetch the complete documentation index at: https://docs.asendia.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Talent AI Search Method
> The Asendia AI sourcing agent uses semantic understanding to match candidates by meaning, not just keywords.
### **Overview**
Semantic search analyzes meaning behind words to uncover the best-fit talent faster — understanding skills, experience, and context beyond simple keyword matches.
### **Data Foundation**
Asendia AI’s sourcing engine is powered by a **global database of over 200 million professional profiles**, aggregated from **public and compliant data sources**.\
Through partnerships with trusted data providers, Asendia AI compiles candidate information from verified sources such as **LinkedIn, public resumes, and professional networks**, enriched with available **contact details (phone number and email)**.
This comprehensive dataset enables the AI to perform deep semantic analysis across massive talent pools, ensuring each search delivers accurate, up-to-date, and actionable results.
***
## **1. Natural Language Input**
Recruiters can describe the role naturally, just as they would explain it to a colleague.\
For example:
> “Find senior full-stack engineers experienced with React, Node.js, and cloud deployment.”
The AI interprets this intent, extracts the relevant entities, such as **role, skills, experience level, and technologies,** and transforms them into structured filters.\
No Boolean syntax is required; the system understands the recruiter’s intent behind the words.
***
## **Extracted Filters**
After the recruiter submits the search, the AI displays a **Filters Tab** that visualizes all extracted criteria, such as:
* **Role:** Full-Stack Engineer
* **Tech Stack:** React, Node.js, AWS
* **Seniority:** Senior
* **Location:** Remote or specific geography
* **Experience:** 5+ years
This helps recruiters validate and adjust the AI’s interpretation before running the full semantic search.
***
## **Filters Page Explained**
The **Filters Page** gives full visibility and control. It’s organized into sections that matter most for engineering searches:
* **Core Role Filters:** Job title, specialization (e.g., Backend, DevOps, Frontend)
* **Technical Skills:** Languages, frameworks, cloud platforms, and tools
* **Experience Filters:** Years of experience, seniority, or domain (e.g., fintech, healthtech)
* **Location & Work Type:** On-site, hybrid, remote, or region-specific
Recruiters can fine-tune or expand these filters, and the AI instantly adapts the sourcing logic.
## **4. AI Matching Process**
The sourcing engine uses a **two-step ranking system** to ensure precision in technical talent matching:
1. **Ranking:** Scores candidates based on direct keyword and metadata relevance (e.g., exact skills, titles, and experience).
2. **Re-ranking:** Applies deep semantic understanding to analyze project descriptions, role context, and related technologies — identifying candidates who may use different terminology but have equivalent experience (e.g., “TypeScript” ≈ “JavaScript,” “GCP” ≈ “AWS”).
This hybrid approach captures hidden talent often missed by Boolean search.
***
## **5. Search Results**
The results show a ranked list of engineering candidates, each with:
* **Fit Score:** AI-calculated probability of match
* **Top Skills & Technologies:** Automatically highlighted from their experience
* **Contextual Summary:** Why they fit the specific role based on projects, frameworks, and industry background
Recruiters can click on profiles to review experience, see AI-generated summaries, and take action directly from their ATS.
***
