Case Study: Smarter Recruitment with Generative AI on Amazon Bedrock

Scalable, Secure CV Analysis for Modern Recruitment Agency

The Client

The recruitment agency specialised in supporting clients across engineering, technology, and energy sectors. With thousands of CVs to process each month, efficiency and accuracy in candidate evaluation are critical to delivering value for clients.

The Challenge

Recruiters at the agency faced a time-consuming and inconsistent CV screening process. Assessing CVs against varied job descriptions required significant manual effort and often lacked a standardised approach. As demand grew, this manual process was becoming a bottleneck, slowing down placements and limiting the ability to scale.

The Solution

Cloud Combinator designed and deployed an AI-powered CV evaluation platform built on AWS. At the core, the solution used Amazon Bedrock with Anthropic’s Claude 3.5 Sonnet to analyse and match CVs against job descriptions and historical placement outcomes.

Recruiters could upload job specifications and candidate CVs via a simple API. The system automatically processed documents stored in Amazon S3, triggered AWS Lambda functions for orchestration, and produced structured scoring outputs (1–5 per candidate). These results were stored in Amazon DynamoDB and made available for recruiter review.

The solution was delivered in a secure, serverless AWS environment with services including:

• Amazon Bedrock for advanced generative AI analysis.

• Amazon S3 for centralised, secure CV and job description storage with cross-region replication.

• AWS Lambda for scalable compute and automation.

• Amazon DynamoDB for state tracking and scoring persistence.

• Amazon VPC with endpoints for network security and governance.

AWS Well-Architected best practices were applied to identity and account governance. IAM Identity Center enabled secure, role-based access with MFA, CloudTrail logging, and SCPs in place. The system was also designed to be maintainable and extensible, providing a foundation for future enhancements to support evolving recruitment needs.

Outcomes

• Reduced CV screening time – what previously took 30 minutes for a handful of CVs could now be completed in minutes.

• Improved consistency – scoring aligned with past successful placements reduced variability in candidate shortlisting.

• Scalable design – recruiters could process high volumes without additional manual overhead.

• Operational resilience – cross-region replication and defined RTO/RPO supported disaster recovery.

• Ongoing optimisation – KPI dashboards, alarms, and runbooks ensure workload health is monitored and improved over time.

Results Summary

The new AWS Bedrock-powered CV analysis solution transformed the recruitment agency's operations. By automating manual screening and embedding consistency into candidate evaluations, the system freed up recruiter time, accelerated placements, and created a secure, scalable platform to support future innovation.