Turning fragmented opportunity discovery into an organized student workflow with AI agents, Mattermost and human–AI collaboration.
1. Introduction
Students come across hundreds of opportunities throughout their academic journey: internships, scholarships, hackathons, fellowships, coding competitions and other career-building programs.
But discovering an opportunity is only the first step.
Students often struggle with finding relevant opportunities, understanding eligibility requirements, keeping track of deadlines, organizing applications and deciding which opportunities deserve their attention.
I wanted to build something that could make this entire process easier.
That idea became OpportunityBridge – Your AI-Powered Student Opportunity Assistant, a project I developed for the Flip the Switch Hackathon 2026, under Track 2: Best Personal / Creative Workflow.
My goal was not simply to create another opportunity-listing website. I wanted to build a personal assistant that could help students discover, evaluate and organize opportunities while using AI agents as collaborators in the discovery workflow.
In this article, I'll share the problem behind the idea, how I built the application, how I integrated Switch and Mattermost, the challenges I encountered and what I learned throughout the process.
2. The Problem I Wanted to Solve
As a computer science student, I understand how difficult it can be to manage academics alongside internships, hackathons, scholarships and other career opportunities.
The problem is not necessarily the lack of opportunities. It is the lack of an organized system to manage them.
Some common challenges include:
- Finding opportunities scattered across different platforms.
- Missing opportunities because of eligibility restrictions.
- Spending time checking requirements that do not match one's profile.
- Forgetting application deadlines.
- Losing track of saved opportunities.
- Managing multiple applications without a clear preparation checklist.
I wanted to bring these activities together into one organized workflow instead of making students manage everything separately.
3. The Solution: OpportunityBridge
OpportunityBridge is a student-focused platform that connects opportunity discovery with personalized matching and application organization.
A student can create a profile containing their degree, branch, academic year, skills and interests.
Based on this information, the application helps students:
- Discover internships, scholarships, hackathons, competitions and fellowships.
- Search and filter opportunities by category and other criteria.
- View eligibility and skill matches based on their profile.
- Understand which matching criteria are met or unmet.
- Save interesting opportunities.
- Track upcoming and urgent deadlines.
- Maintain application preparation checklists with persistent progress.
One important part of the application is its transparent matching system.
OpportunityBridge uses a deterministic, rule-based matching engine to evaluate eligibility and skill relevance. It does not present its matching score as an LLM-generated prediction. This makes the scoring logic easier to understand and explain.
The objective is to reduce the effort spent searching for and organizing opportunities so students can focus more on applying and developing their skills.
4. Why I Chose This Project
I chose this idea because it addresses a problem that is relevant to my own student journey.
Students are constantly encouraged to explore internships, participate in hackathons, apply for scholarships and build their careers. However, managing all these activities together can become overwhelming.
I wanted to create something practical that I would genuinely find useful myself.
I also wanted the project to demonstrate how AI agents can participate in an everyday personal workflow rather than functioning as a chatbot that only answers questions.
That is where Switch became central to the project.
5. Why Switch?
The Flip the Switch Hackathon challenged developers to build workflows where AI agents and humans collaborate through team communication.
I wanted to use Switch as more than an additional integration.
In OpportunityBridge, Switch acts as the orchestration layer connecting the application to specialized agents and their communication environment.
Instead of relying only on manually entering opportunities, students can dispatch fresh discovery requests from the application.
The agents collaborate through Mattermost, conduct opportunity research and prepare structured shortlists. These shortlists can then be synchronized back into OpportunityBridge through the backend.
This makes the project more than a traditional student dashboard. It connects an everyday student task with a coordinated multi-agent workflow.
6. Architecture and Technology
OpportunityBridge uses a frontend, backend, persistence layer and a Switch-powered agent workflow.
Frontend
- React 18 and Vite.
- Tailwind CSS for the user interface.
- Student dashboard and profile.
- Opportunity discovery and filtering.
- Eligibility matches and score breakdowns.
- Saved opportunities and deadline tracking.
Backend
- Node.js and Express.js.
- REST APIs for profile and opportunity management.
- Rule-based matching and eligibility logic.
- Opportunity shortlist validation and synchronization.
- Integration with Switch services.
Switch and Mattermost
- Switch provides multi-agent orchestration.
- Mattermost provides the shared communication environment.
- Specialized agents handle different responsibilities.
- The backend processes the resulting shortlist and stores valid records.
The three specialized agents
1. bridge-coordinator — Task Orchestrator
Coordinates discovery requests, organizes the workflow and delegates research tasks to the opportunity scout.
2. opportunity-scout — Autonomous Web Researcher
Uses the Exa Search API to research opportunities, identify relevant information and prepare structured shortlist JSON files.
3. switch-buddy-admin — Gateway and Webhook Relay
Supports communication between the Express backend and Mattermost, including message delivery and room-history retrieval.
Together, these components connect student requests, agent research, communication and application updates.
7. The Complete Human–AI Workflow
One of the most important parts of OpportunityBridge is the interaction between the student, application, AI agents and Mattermost.
Here is how the workflow operates.
Step 1: Student creates a profile
The student enters academic information, skills and interests.
This information helps OpportunityBridge understand what kinds of opportunities may be relevant.
Step 2: Student requests fresh opportunities
From the application, the student can dispatch a discovery request.
For example:
I am a second-year computer science student interested in cybersecurity internships and technology competitions. Find relevant opportunities.
Step 3: Switch receives the request
The backend normalizes the request, checks for duplicates and forwards a new discovery task to the Switch workflow.
The bridge-coordinator receives the task and coordinates the discovery process with the relevant agent.
Step 4: Agents collaborate through Mattermost
Mattermost provides a shared communication environment for the workflow.
The coordinator and specialized agents exchange messages about the request and discovery task.
The opportunity-scout researches relevant opportunities using Exa and prepares a structured shortlist.
The important idea is that the agents are not isolated tools. Their work is coordinated through a shared communication environment.
Step 5: Shortlist returns to OpportunityBridge
The scout writes the discovered opportunities into a structured JSON file.
The backend then reads the shortlist, checks the required fields, URL format and deadline information, and prevents duplicate records from being imported.
This validation protects the application's data structure. It does not independently guarantee that every external opportunity or website is genuine.
Students should always confirm application details on the original program's official website before applying.
Step 6: Student explores the results
The synchronized opportunities become available in the discovery interface.
The student can then evaluate eligibility, save opportunities, monitor deadlines and prepare applications.
The complete workflow
flowchart TD A["Student Request"] --> B["OpportunityBridge Frontend"] B --> C["Express Backend"] C --> D["Switch"] D --> E["bridge-coordinator"] E --> F["opportunity-scout"] F --> G["Exa Web Research"] F --> H["Mattermost Communication"] F --> I["Structured Shortlist JSON"] I --> J["Backend Validation"] J --> K["JSON Persistence Layer"] K --> L["Updated Opportunities in UI"] L --> M["Student Reviews and Applies"]Student → OpportunityBridge → Switch → AI Agents → Mattermost → Backend Validation → Database → Student
8. Mattermost: Bringing Human–AI Collaboration Together
An important part of this workflow is human–AI collaboration.
Mattermost acts as the shared communication space where team members with access to the same configured workspace can interact with the agent workflow.
For example, a teammate could contribute a request such as:
Find upcoming frontend development internships suitable for second-year students.
Another teammate could ask the coordinator to explore scholarship opportunities or review a particular discovery task.
Team members can observe agent communication, contribute requests and follow the workflow instead of relying entirely on separate, disconnected AI conversations.
In the current project setup, the Mattermost workspace and Switch services run locally. Additional teammates need access to the same configured workspace and running services to participate. Remote shared access is an area for further improvement.
This is the collaboration model I designed OpportunityBridge to support.
9. What Makes OpportunityBridge a Personal / Creative Workflow?
Track 2 focuses on personal and creative workflows.
For me, OpportunityBridge represents a practical personal productivity workflow built around a real student need.
It brings together multiple activities that students normally manage separately:
- Discovering opportunities.
- Evaluating personal eligibility.
- Organizing saved opportunities.
- Monitoring deadlines.
- Preparing applications.
- Delegating fresh opportunity discovery to AI agents.
The creative aspect is the combination of a personalized student profile, structured opportunity management and a collaborative AI-agent workflow.
Switch is not just mentioned in the project. It connects the discovery process with specialized agents and Mattermost communication.
Without the Switch-powered workflow, the application would lose its agent-coordinated fresh opportunity discovery capability.
That makes Switch an important part of the project's architecture and purpose.
10. Challenges and What I Learned
Building OpportunityBridge taught me several things beyond frontend and backend development.
Some of the major challenges included:
Agent orchestration: Connecting an application to multiple specialized agents required understanding how requests move through the Switch workflow.
Service integration: Running Switch, Mattermost and the application together required managing multiple services and their configurations.
Data synchronization: Agent-generated opportunity records needed validation and duplicate handling before being imported into the application.
Persistence: Saved opportunities and application checklist progress needed to remain available after refreshing or reopening the application.
Workflow design: I learned that a useful AI workflow is not simply about generating responses. It is about connecting agents to real tasks and producing results that users can act on.
Working through these challenges helped me better understand the practical value of agent orchestration, backend integration and human–AI collaboration.
11. Current Implementation and Future Scope
The current implementation includes:
- Student profile management.
- Opportunity discovery and filtering.
- Rule-based eligibility and skill matching.
- Saved opportunities.
- Deadline tracking.
- Persistent application checklists.
- Switch-powered agent dispatch.
- Mattermost agent communication.
- Opportunity shortlist synchronization.
There are also areas I would like to improve further:
- Secure shared access for remote teammates.
- More comprehensive opportunity source verification.
- More advanced personalized recommendations.
- Additional application progress tracking.
- Improved deployment and multi-user support.
I see OpportunityBridge as a foundation that can grow into a more comprehensive student opportunity management platform.
12. Conclusion
OpportunityBridge started with a simple question:
What if students had one place to discover opportunities, understand their eligibility and organize their applications, while AI agents helped with the discovery process?
Through this project, I explored how Switch can turn specialized AI agents into collaborators within a practical personal workflow.
The most valuable lesson for me was that AI becomes more useful when it is connected to a real task, a communication environment and an application where its work produces actionable results.
I built OpportunityBridge to solve a problem I understand personally, and this hackathon gave me an opportunity to explore how human–AI collaboration can make that workflow better.
This project has been a learning experience in building, integrating, testing and experimenting with tools beyond my usual development workflow.
Built with curiosity, experimentation and a student-first mindset.
Project Links
- GitHub Repository: https://github.com/pihush09/OpportunityBridge
- Demo Video: https://youtu.be/tjhb7nR2F44
Hackathon
Built for the Flip the Switch Hackathon 2026, organized by WeMakeDevs and DevTools Academy, under Track 2 – Best Personal / Creative Workflow.
