RecruSmart — AI Recruitment Platform
A microservices platform that matches CVs to job offers with machine learning and processes applications asynchronously.
Overview
RecruSmart is a recruitment platform built as a set of FastAPI microservices. Candidates and offers are stored in MongoDB, a machine-learning service scores how well each CV matches an offer, and RabbitMQ handles the asynchronous processing between services. Everything runs in Docker and Kubernetes.
The problem
Recruiters spend a lot of time reading CVs that do not match the offer, and a single monolithic application would struggle when many applications arrive at once.
What I built
I split the platform into services (candidates, offers, matching, notifications) exposed as REST APIs with FastAPI. When a CV is uploaded, a message is published on RabbitMQ; the matching service extracts features, computes a relevance score against each offer with a machine-learning model and stores the ranking in MongoDB. Services are containerised and deployed on Kubernetes.
Results
- Automatic ranking of candidates for each offer
- Services that scale independently thanks to the message queue
- A reproducible deployment with Docker and Kubernetes manifests
Highlights
- Microservices architecture with FastAPI
- Asynchronous processing with RabbitMQ
- ML-based CV ↔ offer matching and ranking
- MongoDB storage, Docker and Kubernetes deployment
