AIME — Agentic Recruitment Platform
End-to-end agentic RAG ecosystem that automates global candidate sourcing, standardization, and matching at scale.
- LangChain
- LangGraph
- RAG
- FAISS
- AWS
- MongoDB
- FastAPI
- Python
Overview
Traditional recruitment at global scale is manual, slow, and inconsistent. Sourcing candidates across multiple countries, parsing CVs in different formats, and matching them to job descriptions took days of human effort with no standardization. AIME is a production agentic RAG ecosystem that automates the entire pipeline — from CV ingestion and standardization to semantic matching and job board integration. As Lead System Architect, I designed the full architecture and led a team of two developers from design to production on AWS EC2. An intelligent ingestion engine monitors a SharePoint Data Lake, using LLMs to detect candidate locations from CV metadata and route files to country-specific S3 buckets and MongoDB collections. A preprocessing layer standardizes every CV into structured summaries and skills taxonomies. The Agentic CV Matcher then performs deep semantic search using RAG over FAISS, dynamically selecting the right geographic context per job description. Integration with 4 global job boards completes the loop — automating ad creation and harvesting external candidates into a unified intelligence hub.
Architecture
Results
- 10,000+ CVs processed and standardized autonomously
- Multi-day manual recruitment process reduced to minutes
- Integrated with 4 global job boards for automated candidate sourcing
- Led a team of 2 developers end-to-end from design to production
- Deployed on AWS EC2 with MongoDB, S3, and FAISS vector database