10+ Years In Business | 4 Continents |
16+ Countries | 32+ Locations

Contact Us

Redefining Candidate Matching with 
Graph RAG and Multi-Agent AI for RightSourceAI Test

Project Overview

A concise overview of how RightSourceAI leverages next-gen RAG techniques; specifically Graph RAG and multi-agent architectures; to solve one of the most pressing challenges in recruitment: matching accuracy. With a sharp focus on speed, precision, and cost-efficiency, the platform delivers a powerful tool for engineering teams and recruiters in fast-scaling SaaS environments.

Test

Industry

  •  Recruitment tech test

Initial Go-to-
Market Timeline

  • 3 months test

Tech Stack

  • Graph RAG (graph-based retrieval-augmented generation),
  • AWS Neptune (graph DB),
  • AWS RDS,
  • Bedrock (LLM orchestration), test
  • Multi-agent orchestration,
  • LangChain, LangGraph,
  • LangSmith,
  • NextJs,
  • AWS Fargate,
  • SSO via Google Authentication test

Team
Composition

  • 1 Product Manager
  • 1 Frontend Engineer
  • 1 Backend Engineer test
  • 1 Data Scientist & LLMOps
  • 1 QA Analyst test

Client Background

RightSourceAI was developed to address one of the most persistent and costly challenges in technical hiring: slow screening cycles driven by inaccurate candidate matching.

Most traditional applicant tracking and sourcing systems rely on keyword matching or vector similarity. While these approaches can surface large volumes of candidates, they introduce two critical problems:

  1. High false positives – recruiters and hiring managers must manually screen many irrelevant profiles
  2. High false negatives – qualified candidates are missed because resumes do not contain exact keywords

These issues compound into longer hiring cycles, higher recruiter costs, and excessive time spent by engineering leaders in interviews that should never have happened.

Originally built to support Athenaworks’ internal recruiting operations, RightSourceAI was designed to scale across both SaaS and enterprise organizations hiring for complex engineering, data, and product roles. test

The Challenge

Recruiting and engineering teams face two compounding issues:

  • Too many weak matches from keyword/vector-based systems
  • Delayed discovery of strong candidates, especially for specialized roles

Even modern AI-powered recruiting tools struggle to deliver context-rich matches at speed. For time-sensitive technical hiring, this leads to:

  • Increased recruiter screening hours
  • Lost productivity from engineering interview loops
  • Slower delivery against product and revenue goals

Even modern AI-powered recruiting tools struggle to deliver context-rich matches at speed. For time-sensitive technical hiring, this leads to:

  • Increased recruiter screening hours
  • Lost productivity from engineering interview loops
  • Slower delivery against product and revenue goals test

The Solution:

RightSource introduces a new candidate-matching architecture built on:

  • Graph RAG: Uses graph databases (AWS Neptune) to model relationships between skills, roles, experience, and requirements
  • Multi-agent AI: Parallelizes resume analysis and matching tasks to deliver near real-time results
  • Cost-aware LLM orchestration: Selects models via AWS Bedrock to balance accuracy, latency, and inference cost

Together, this enables deep semantic matching at production scale.

Key Capabilities

  • Upload or paste a job description
  • Instantly return AI-matched candidates from a pool of 25,000+
  • Secure, SSO-enabled access (Google login for MVP)

Optional integration with Athenaworks’ Gatekeeper service for curated, pre-interviewed candidates test

Download PDF
Contact Us!
Linkedin-in X-twitter Instagram

©2025 CANWILL ATHENAWORKS HOLDINGS LLC