🏆 2nd Place — AWS hackathon
Nole Path is an AI-powered career readiness platform designed to guide students from skill discovery to internship placement by integrating resume optimization, job matching, mentorship recommendations, and opportunity sourcing into a unified system. The prototype explored scalable, cloud-native architecture for delivering personalized guidance using agent-style workflows and AWS infrastructure.
This project was developed as part of the AWS Design Sprint competition and focused on demonstrating how generative AI and serverless services can be combined to support large-scale student career services.
A deployed prototype of the platform can be accessed here:
🔗 https://nolepath.replit.app/dashboard
Note: This demo reflects the competition prototype and may not include full production functionality.
Students often struggle to identify relevant internships, understand employer skill expectations, and translate academic experience into competitive applications. At the same time, career centers face scalability challenges in providing personalized advising.
Nole Path addresses this by:
- Building dynamic student skill profiles
- Matching students to opportunities using compatibility scoring
- Providing AI-driven resume and preparation feedback
- Aggregating internship listings from multiple sources
- AI career coaching and skill gap identification
- Resume alignment and optimization guidance
- Internship compatibility scoring
- Opportunity aggregation via scraping workflows
- Mentorship and preparation recommendations
- Scalable architecture supporting large user bases
The prototype leveraged AWS services to simulate a modular agent-driven pipeline:
- Amazon Bedrock — LLM-driven coaching and interaction
- AWS Lambda — Serverless compute for workflow execution
- API Gateway — Endpoint routing and orchestration
- Amazon S3 — Storage for resumes, job data, and profiles
- Amazon Cognito — Authentication and user identity
- AWS IAM — Access control and permissions
- Amazon CloudWatch — Monitoring and logging
This stack enabled a scalable, event-driven architecture capable of handling dynamic user interactions.
Prototype analysis projected infrastructure costs of approximately:
~$60/day for 1,000 users
Demonstrating feasibility of scalable deployment in an academic environment.
As part of a collaborative team effort, I contributed across technical and design components including:
- Participating in system architecture planning and workflow design
- Assisting with prototype development and feature refinement
- Contributing to technical documentation and presentation materials
- Communicating system functionality and value during judging
My role emphasized cross-functional collaboration between technical implementation, system planning, and communication.
- 🥈 2nd Place — AWS Design Sprint
- Demonstrated scalable AI-assisted advising concept
- Validated feasibility of agent-style architecture in higher education workflows
- Received positive technical and strategic feedback from judges
Due to competition and team ownership constraints, source code is not publicly available.
This repository documents the architecture, design rationale, and project scope for professional and academic review.
Ermithe Tilusca
Computer Science & Applied Mathematics
Florida State University