NextStep: AI-Powered Student Opportunity Finder by Hari
Students often struggle to discover internships, hackathons, and competitions that match their skills. NextStep aims to bridge this gap by analyzing student skills, identifying relevant opportunities, and explaining both direct and transferable-skill eligibility.
NextStep: Building a Student Opportunity Finder with AWS
Finding the right internship, hackathon, competition, or other student opportunity can be surprisingly difficult.
Students often have to search across multiple websites, read different eligibility requirements, compare their skills with those requirements, and decide whether an opportunity is actually suitable for them.
I wanted to build something that could make this process simpler.
That idea became NextStep — Student Opportunity Finder.
The Problem
The internet has no shortage of opportunities for students. The challenge is discovering the opportunities that are actually relevant to an individual student.
A student may have skills in C++, Java, Python, data structures, algorithms, object-oriented programming, web development, or other areas, but finding opportunities that match those skills can require a lot of manual searching.
There is another problem with simple keyword-based eligibility checking.
For example, imagine a hackathon asks for Python, while a student has strong C++, OOP, DSA, and algorithm skills.
A basic system might simply say that the student does not match the requirement.
But the student's existing programming knowledge can still be valuable.
This led me to the idea behind NextStep: instead of only asking "Does this student have the exact keyword?", the system should also help answer "What existing skills can this student build upon?"
The Solution
NextStep is a web application that helps students discover opportunities based on their skills and interests.
The application analyzes student information against opportunity requirements and provides relevant results.
One of the main ideas behind the project is distinguishing between two types of eligibility:
Direct Eligibility
An opportunity can be a direct match when the student's existing skills satisfy the relevant requirements.
For example, if an opportunity requires Java and the student already has Java experience, the system can identify that direct connection.
Transferable-Skills Eligibility
The second category is designed for situations where a student does not have an exact skill match but has related technical foundations.
For example, a student with strong C++, object-oriented programming, data structures, and algorithms knowledge may not know Python yet.
Instead of treating the student as simply unsuitable, NextStep can identify the transferable foundation and explain the missing skill that the student would need to develop.
The goal is not to falsely tell a student that they are eligible.
The goal is to help them understand their current position and the path they can take toward an opportunity.
How NextStep Works
The basic user journey is:
Student → Skills and Interests → Opportunity Search → Eligibility Analysis → Relevant Results
The student provides their skills and interests through the application.
NextStep then processes that information and compares it with the requirements of available opportunities.
The system can return opportunities along with relevant eligibility information, helping the student understand why an opportunity may be useful to them.
The project also uses a curated opportunity catalogue so that the search process is based on structured opportunity information rather than simply producing unrelated web results.
Technology and Architecture
NextStep is built as a web application with a cloud-based backend.
The deployed architecture follows this flow:
Browser → Amazon CloudFront → Amazon S3 → API Gateway → AWS Lambda
The frontend is a Vite-based application hosted in an Amazon S3 bucket.
Amazon CloudFront provides the distribution layer for delivering the frontend.
The frontend communicates with the backend through an Amazon API Gateway HTTP API.
AWS Lambda handles the server-side functionality, including search, opportunity information, and eligibility-related operations.
NextStep also contains an optional AI-agent capability. A separate Python-based Lambda function runs a Strands Agent, which can interact with the search functionality and use Amazon Bedrock for model inference.
This architecture separates the user interface, API layer, backend processing, and AI capabilities while keeping the overall application relatively lightweight.
How I Used AWS
AWS was used as the cloud foundation for deploying NextStep.
Amazon S3
Amazon S3 is used to store the static frontend application.
Instead of requiring a traditional web server just to serve the frontend files, the application can be hosted as static content in S3.
Amazon CloudFront
CloudFront is used to distribute the frontend application.
The S3 bucket is kept private, while CloudFront acts as the public distribution layer.
This provides a cleaner separation between the stored frontend files and the public-facing application.
Amazon API Gateway
API Gateway provides the HTTP API through which the frontend communicates with the backend.
The API handles requests such as searching opportunities, retrieving opportunity information, and performing eligibility-related operations.
AWS Lambda
AWS Lambda runs the backend functionality without requiring a continuously running server.
The search functionality runs in a Node.js 22 Lambda function, while the optional agent functionality runs in a Python 3.12 Lambda function.
This serverless approach allowed me to build the backend around individual functions rather than maintaining a traditional always-running server.
Amazon Bedrock
Amazon Bedrock is used for the optional AI-agent capability.
The Strands Agent runs inside Lambda and can use a Bedrock model for AI inference.
The agent can also interact with the application's search functionality, allowing the AI layer to work with the same opportunity data used by the application.
AWS SAM
AWS Serverless Application Model (SAM) is used to define the AWS infrastructure as code.
The project's
template.yaml describes the required cloud resources and their configuration.This made the deployment process more reproducible than manually creating every AWS resource through the console.
Why I Chose This Architecture
I wanted the architecture to remain practical for the current scale of the project.
A student opportunity finder does not necessarily need a large and complicated infrastructure from the beginning.
The combination of S3, CloudFront, API Gateway, Lambda, and Bedrock gives NextStep a serverless foundation while keeping the different parts of the application separated.
The repository also contains an OpenSearch component that I experimented with during development. However, OpenSearch is not part of the current production deployment architecture.
For the current version, I use a curated opportunity catalogue together with application-level search and eligibility logic. This keeps the deployed system simpler and avoids adding infrastructure that is not necessary for the current project scope.
What I Learned
Building NextStep taught me that developing a project is much more than writing code.
I had to think about the actual problem a student faces, design a user journey, structure opportunity data, implement matching and eligibility logic, work with AI capabilities, and understand how a locally developed application can be deployed using cloud services.
Working with AWS also gave me practical experience with serverless architecture, infrastructure as code, API design, Lambda functions, cloud deployment, and permissions.
One of the biggest lessons for me was that a good technical solution should remain focused on the user's actual problem.
It is easy to add technologies simply because they are available. It is much harder to decide which technologies are actually useful for the product.
Future Improvements
There are several areas where I would like to take NextStep further.
Future versions could include:
- More opportunity sources
- More sophisticated personalization
- Improved skill-gap analysis
- Additional language support
- Opportunity deadline tracking
- Notifications for relevant opportunities
- More advanced AI-assisted recommendations
- Larger and continuously updated opportunity datasets
Conclusion
NextStep started with a simple observation: opportunities exist everywhere, but students can struggle to find the ones that are relevant to them.
I wanted to build a system that does more than simply return a list of search results.
By combining opportunity discovery, direct eligibility, transferable-skill analysis, serverless AWS infrastructure, and an optional AI-agent layer, NextStep aims to help students understand not only what opportunities are available, but also what they can do next to reach them.
That is the idea behind NextStep:
Find opportunities. Understand your potential. Take your next step.
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article