A comprehensive, intelligent job search assistant that empowers students with personalized job recommendations, resume-based career insights, and automated daily notifications — built using AWS Bedrock AgentCore, Nova Pro, and event-driven serverless architecture.

The Agentic Job Search Assistant leverages AWS Bedrock’s agentic capabilities to automate the end-to-end student job search process — from resume parsing to daily career recommendations.

It integrates real-time user interactions (via live search) with automated batch processes (for job matching and communication), providing a seamless and intelligent job discovery experience.

• Multi-Agent System (AgentCore):
  Includes a Routing Agent, Career Exploration Agent, and Job Search Agent for modular and dynamic decision-making.
• AI-Powered Resume Parsing:
  Extracts key entities and skills using Nova Pro and stores structured student profiles in DynamoDB.
• Automated Notifications:
  Sends daily job updates via AWS SES (email) and SNS (SMS).
• Event-Driven Architecture:
  Uses Amazon EventBridge and SQS for time-triggered and asynchronous job processing.
• Knowledge Graph + RAG:
  Combines Bedrock Knowledge Base, Graph RAG, and Neptune Graph for context-aware job matching and career exploration.
• Secure File Storage:
  User resumes and resources are securely managed in Amazon S3.

The system is divided into three main modules:

• Students upload resumes → stored in Amazon S3.
• Lambda + Nova Pro parses resumes and extracts skills.
• Processed profiles are stored via API Gateway → DynamoDB.
• AgentCore runtime routes user queries to the appropriate agent:
  • Career Exploration Agent: Provides insights and skill-based recommendations.
  • Job Search Agent: Fetches job postings and returns AI-filtered results.

• Triggered daily via Amazon EventBridge.
• Lambda adds new student job search tasks to Amazon SQS queue.
• Another Lambda processes queued tasks to perform job searches and updates DynamoDB with results.

• Triggered daily at configurable times (e.g., 9 AM).
• Fetches new job recommendations from DynamoDB.
• Sends personalized email and SMS notifications using:
  • Amazon SES (Email Service)
  • Amazon SNS (SMS Service)

• Bedrock Knowledge Base connects with:
  • S3 Vector Store: Stores semantic embeddings of career resources.
  • Neptune Graph: Stores relational data between jobs, skills, and industries.
  • Graph RAG: Enhances contextual retrieval for AI-driven responses.

Follow the complete deployment guide here:
👉 docs/DEPLOYMENT.MD

Includes setup for:

• Lambda functions
• API Gateway routes
• Bedrock AgentCore configuration
• DynamoDB and Neptune initialization
• SQS and EventBridge scheduling
• SES/SNS permissions for notifications

After successful deployment:

1. Upload student resumes via the web interface.
2. Use Live Search for real-time job queries.
3. Receive daily job recommendations via email and SMS.
4. Explore personalized career resources powered by Graph RAG.

For detailed steps → Post Deployment Setup

The entire system is serverless and event-driven, leveraging:

• Compute: AWS Lambda
• Storage: S3, DynamoDB
• AI Models: Bedrock (Claude Sonnet, Nova Pro)
• Messaging: SQS, SNS, SES
• Graph & RAG: Neptune Graph, Bedrock Knowledge Base
• Automation: EventBridge

See full infrastructure details in docs/INFRASTRUCTURE.MD.

You can easily extend or modify:

• LLM Models: Swap Bedrock models (e.g., Nova Pro → Claude 3.5 Sonnet).
• Notification Logic: Add more conditions or new channels (Slack, Teams).
• Agent Behavior: Adjust Routing Agent logic to support new intents.

See docs/modificationGuide.md for instructions.

Dsigned and developed by

• Aryan Khanna
• Aarav Matalia
• Sayantika Paul

Licensed under the MIT License. See the LICENSE file for details.